AB. Delta method.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "47e421ca-a8ac-4df3-aab5-c8c6606e534f"
      },
      "source": [
        "# ДЗ-2 по Теории и практике онлайн экспериментов\n",
        "\n",
        "### Задание 1 (1.5 балла)\n",
        "\n",
        "В приложенном файле `sleep.csv` приведено время засыпания одного и того же человека в разные дни.\n",
        "\n",
        "Нужно обработать эти данные и указать такое время суток (с точностью до минуты), после которого с вероятностью 0.9 этот человек уже спит.\n",
        "\n",
        "Как можно оценить стандартное отклонение этой точечной оценки?"
      ],
      "id": "47e421ca-a8ac-4df3-aab5-c8c6606e534f"
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "wBJIOjGHPKJd"
      },
      "outputs": [],
      "source": [
        "# !pip install gdown"
      ],
      "id": "wBJIOjGHPKJd"
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "id": "cc52bca6"
      },
      "outputs": [],
      "source": [
        "import pandas as pd\n",
        "import requests\n",
        "from io import StringIO\n",
        "\n",
        "# Прямая ссылка на скачивание файла\n",
        "url = 'https://drive.google.com/uc?id=1EhaU8BcM9vxkVkmOycolype77B1yzEPr'\n",
        "\n",
        "# Загрузка файла\n",
        "response = requests.get(url)\n",
        "\n",
        "# Преобразование данных в строку\n",
        "data = response.text\n",
        "\n",
        "# Чтение данных CSV\n",
        "df = pd.read_csv(StringIO(data))"
      ],
      "id": "cc52bca6"
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "id": "zwambNXHo36P",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "faf26289-5390-47f1-8c51-40a545721675"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "            Unnamed: 0  indicator  time_offset\n",
              "0  2017-12-01 21:00:00          0          -60\n",
              "1  2017-12-01 21:01:00          0          -59\n",
              "2  2017-12-01 21:02:00          0          -58\n",
              "3  2017-12-01 21:03:00          0          -57\n",
              "4  2017-12-01 21:04:00          0          -56"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-3fc70dc9-105b-4dc9-8a03-7815e2e9ee2d\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Unnamed: 0</th>\n",
              "      <th>indicator</th>\n",
              "      <th>time_offset</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>2017-12-01 21:00:00</td>\n",
              "      <td>0</td>\n",
              "      <td>-60</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2017-12-01 21:01:00</td>\n",
              "      <td>0</td>\n",
              "      <td>-59</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2017-12-01 21:02:00</td>\n",
              "      <td>0</td>\n",
              "      <td>-58</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>2017-12-01 21:03:00</td>\n",
              "      <td>0</td>\n",
              "      <td>-57</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2017-12-01 21:04:00</td>\n",
              "      <td>0</td>\n",
              "      <td>-56</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-3fc70dc9-105b-4dc9-8a03-7815e2e9ee2d')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-3fc70dc9-105b-4dc9-8a03-7815e2e9ee2d button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-3fc70dc9-105b-4dc9-8a03-7815e2e9ee2d');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-390e3b58-61c8-492d-9c63-a540bf2fda5c\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-390e3b58-61c8-492d-9c63-a540bf2fda5c')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-390e3b58-61c8-492d-9c63-a540bf2fda5c button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "df",
              "summary": "{\n  \"name\": \"df\",\n  \"rows\": 11340,\n  \"fields\": [\n    {\n      \"column\": \"Unnamed: 0\",\n      \"properties\": {\n        \"dtype\": \"object\",\n        \"num_unique_values\": 11340,\n        \"samples\": [\n          \"2017-12-08 23:16:00\",\n          \"2017-12-04 21:03:00\",\n          \"2018-01-25 21:02:00\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"indicator\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0,\n        \"min\": 0,\n        \"max\": 1,\n        \"num_unique_values\": 2,\n        \"samples\": [\n          1,\n          0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"time_offset\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 51,\n        \"min\": -60,\n        \"max\": 119,\n        \"num_unique_values\": 180,\n        \"samples\": [\n          -41,\n          -18\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 3
        }
      ],
      "source": [
        "df.head()"
      ],
      "id": "zwambNXHo36P"
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "6l0zh253vDLx"
      },
      "outputs": [],
      "source": [
        "df_sorted = df.sort_values(by='time_offset') # отсортируем по возрастанию"
      ],
      "id": "6l0zh253vDLx"
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "3BpkbahKv73B",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "61a27bcf-f074-4e07-8219-6bd61ef94eb9"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                Unnamed: 0  indicator  time_offset\n",
              "0      2017-12-01 21:00:00          0          -60\n",
              "4320   2017-12-25 21:00:00          0          -60\n",
              "1260   2017-12-08 21:00:00          0          -60\n",
              "10080  2018-01-26 21:00:00          0          -60\n",
              "6840   2018-01-08 21:00:00          0          -60"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-a36c6837-1f9b-4b9a-8231-2c2f505cb090\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Unnamed: 0</th>\n",
              "      <th>indicator</th>\n",
              "      <th>time_offset</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>2017-12-01 21:00:00</td>\n",
              "      <td>0</td>\n",
              "      <td>-60</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4320</th>\n",
              "      <td>2017-12-25 21:00:00</td>\n",
              "      <td>0</td>\n",
              "      <td>-60</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1260</th>\n",
              "      <td>2017-12-08 21:00:00</td>\n",
              "      <td>0</td>\n",
              "      <td>-60</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10080</th>\n",
              "      <td>2018-01-26 21:00:00</td>\n",
              "      <td>0</td>\n",
              "      <td>-60</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6840</th>\n",
              "      <td>2018-01-08 21:00:00</td>\n",
              "      <td>0</td>\n",
              "      <td>-60</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-a36c6837-1f9b-4b9a-8231-2c2f505cb090')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-a36c6837-1f9b-4b9a-8231-2c2f505cb090 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-a36c6837-1f9b-4b9a-8231-2c2f505cb090');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-59a5077e-476f-4e16-9728-8b89c93e3f3c\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-59a5077e-476f-4e16-9728-8b89c93e3f3c')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-59a5077e-476f-4e16-9728-8b89c93e3f3c button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "df_sorted",
              "summary": "{\n  \"name\": \"df_sorted\",\n  \"rows\": 11340,\n  \"fields\": [\n    {\n      \"column\": \"Unnamed: 0\",\n      \"properties\": {\n        \"dtype\": \"object\",\n        \"num_unique_values\": 11340,\n        \"samples\": [\n          \"2017-12-07 21:22:00\",\n          \"2018-01-13 21:08:00\",\n          \"2018-01-09 23:37:00\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"indicator\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0,\n        \"min\": 0,\n        \"max\": 1,\n        \"num_unique_values\": 2,\n        \"samples\": [\n          1,\n          0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"time_offset\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 51,\n        \"min\": -60,\n        \"max\": 119,\n        \"num_unique_values\": 180,\n        \"samples\": [\n          -41,\n          -18\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 5
        }
      ],
      "source": [
        "df_sorted.head()"
      ],
      "id": "3BpkbahKv73B"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gRTKCnnBRBga"
      },
      "source": [
        "Мы видим, что у нас в таблице в столбце `indicator` указаны значения, когда человек спит и когда не спит, в столбце `time_offset` указаны значения в минутах в расчете от 22:00:00. Поскольку нам надо найти время суток в минутах, после которого человек с вероятностью 0,9 будет спать, то будем искать квантиль 0,9 случайной величины \"человек спит\". Для этого оставим столбец только, где человек спит (соответствует $indicator==1$), отсортируем наблюдения в порядке возрастания и найдем квантиль отранжированного ряда, полученное число - это время суток, после которого человек с вероятностью 0,9 будет спать (иначе говоря, 90% наблюдений, когда человек спит будут лежать левее этого значения). В конце просто переведем это значение в сутки. Оценку стандартного отклонения найдем с помощью бутстрапа."
      ],
      "id": "gRTKCnnBRBga"
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "BuPTbIbiwegR",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d3c25b5b-e87c-415e-b278-104d69c59226"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Оценка 90% квантиля: 109.00\n",
            "Оценка std для 90% квантиля: 0.45\n"
          ]
        }
      ],
      "source": [
        "import numpy as np\n",
        "\n",
        "sleep = df[df['indicator'] == 1]['time_offset']  # оставим только таблицу с заснувшими людьми\n",
        "\n",
        "n = len(sleep)            # размер исходной выборки\n",
        "B = n            # количество генерируемых подвыборок\n",
        "\n",
        "quantile = np.quantile(sleep, 0.9)\n",
        "bootstrap_quantiles = []\n",
        "for _ in range(B):\n",
        "    bootstrap_values = np.random.choice(sleep, n, True)\n",
        "    bootstrap_quantiles.append(np.quantile(bootstrap_values, 0.9))\n",
        "std = np.std(bootstrap_quantiles)\n",
        "print(f'Оценка 90% квантиля: {quantile:0.2f}')\n",
        "print(f'Оценка std для 90% квантиля: {std:0.2f}')"
      ],
      "id": "BuPTbIbiwegR"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ux3KYX2XfQSY"
      },
      "source": [
        "### 109 минут соответствует времени **23:49:00**. Стандартное отклонение точечной оценки квантиля $+-$ 0.44 мин примерно."
      ],
      "id": "ux3KYX2XfQSY"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2f15c6a1"
      },
      "source": [
        "-----------\n",
        "Представим, что у нас есть сервис для продажи некоторых товаров (например, мыла). Мы выкатываем новые фичи и проводим тесты, а затем сохраняем логи каждого теста в следующем формате:\n",
        "\n",
        "- `user_id` - уникальный id посетителя сайта\n",
        "- `dt` - дата события\n",
        "- `event` - название события\n",
        "- `group` - группа, в которую попал пользователь (распределение может происходить на разных этапах воронки для разных тестов - например, при первом посещении или при первом добавлении в корзину)\n",
        "- `visit_id` - уникальный id визита\n",
        "- `cart_id` - уникальный id корзины\n",
        "- amount - текущая сумма корзины\n",
        "- `num_items` - текущее количество товаров в корзине"
      ],
      "id": "2f15c6a1"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9e50e1aa-da1c-47d6-9b77-00a6d50c37d8"
      },
      "source": [
        "### Задание 2 (0.5 балла)\n",
        "\n",
        "Мы сделали рекомендательную систему, которая после добавления товара в корзину предлагает добавить похожие или комплементарные. Логи проведенного теста сохранены в файле `ab_test_logs.pkl`. Проверьте гипотезу о том, что в тестовой группе для 70% пользователей выросла выручка (revenue per user)."
      ],
      "id": "9e50e1aa-da1c-47d6-9b77-00a6d50c37d8"
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "id": "sUofM48zgcKj"
      },
      "outputs": [],
      "source": [
        "import requests\n",
        "import pickle\n",
        "\n",
        "# Прямая ссылка на скачивание файла\n",
        "url = 'https://drive.google.com/uc?id=1J6RpGua4KoN87wuLhzAUO1j-Ml-rV54S'\n",
        "\n",
        "# Загрузка файла\n",
        "response = requests.get(url)\n",
        "\n",
        "# Чтение данных pickle\n",
        "df1 = pickle.loads(response.content)"
      ],
      "id": "sUofM48zgcKj"
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "id": "4LVIemy8-MRr",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "556d4558-163f-43c6-f592-73b129cf0b77"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "  user_id         dt          event group              visit_id cart_id  \\\n",
              "0  3Y6FR5 2024-01-17  first-visited   NaN                   NaN     NaN   \n",
              "1  6W79RT 2024-01-17  first-visited   NaN                   NaN     NaN   \n",
              "2  6W79RT 2024-01-17     registered  test                   NaN     NaN   \n",
              "3  6W79RT 2024-01-17     shop-visit  test  3qzhu3sar2vee6rzssyr     NaN   \n",
              "4  6W79RT 2024-01-18     shop-visit  test  siglppuqke79ts5a5r4t     NaN   \n",
              "\n",
              "   amount  num_items  \n",
              "0     NaN        NaN  \n",
              "1     NaN        NaN  \n",
              "2     NaN        NaN  \n",
              "3     NaN        NaN  \n",
              "4     NaN        NaN  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-89dc358f-2002-407a-b79a-2156640eb62f\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user_id</th>\n",
              "      <th>dt</th>\n",
              "      <th>event</th>\n",
              "      <th>group</th>\n",
              "      <th>visit_id</th>\n",
              "      <th>cart_id</th>\n",
              "      <th>amount</th>\n",
              "      <th>num_items</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>3Y6FR5</td>\n",
              "      <td>2024-01-17</td>\n",
              "      <td>first-visited</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>6W79RT</td>\n",
              "      <td>2024-01-17</td>\n",
              "      <td>first-visited</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>6W79RT</td>\n",
              "      <td>2024-01-17</td>\n",
              "      <td>registered</td>\n",
              "      <td>test</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>6W79RT</td>\n",
              "      <td>2024-01-17</td>\n",
              "      <td>shop-visit</td>\n",
              "      <td>test</td>\n",
              "      <td>3qzhu3sar2vee6rzssyr</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>6W79RT</td>\n",
              "      <td>2024-01-18</td>\n",
              "      <td>shop-visit</td>\n",
              "      <td>test</td>\n",
              "      <td>siglppuqke79ts5a5r4t</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-89dc358f-2002-407a-b79a-2156640eb62f')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-89dc358f-2002-407a-b79a-2156640eb62f button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-89dc358f-2002-407a-b79a-2156640eb62f');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-8f72d7c2-8ce0-492c-856d-87e6876c32c1\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-8f72d7c2-8ce0-492c-856d-87e6876c32c1')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-8f72d7c2-8ce0-492c-856d-87e6876c32c1 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "df1",
              "summary": "{\n  \"name\": \"df1\",\n  \"rows\": 51382,\n  \"fields\": [\n    {\n      \"column\": \"user_id\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 9054,\n        \"samples\": [\n          \"3Q129M\",\n          \"BHMY36\",\n          \"KUZOTE\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"dt\",\n      \"properties\": {\n        \"dtype\": \"date\",\n        \"min\": \"2024-01-10 00:00:00\",\n        \"max\": \"2024-01-24 00:00:00\",\n        \"num_unique_values\": 15,\n        \"samples\": [\n          \"2024-01-14 00:00:00\",\n          \"2024-01-15 00:00:00\",\n          \"2024-01-17 00:00:00\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"event\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 5,\n        \"samples\": [\n          \"registered\",\n          \"cart-paid\",\n          \"shop-visit\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"group\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 2,\n        \"samples\": [\n          \"control\",\n          \"test\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"visit_id\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 29998,\n        \"samples\": [\n          \"tsha83gw00n4k97nb0la\",\n          \"t012s0ahengc4or9smj4\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"cart_id\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 2757,\n        \"samples\": [\n          \"os3dmm6spj\",\n          \"aoj7bika3y\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"amount\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 3958.4343152104493,\n        \"min\": 26.63452376890258,\n        \"max\": 55109.90049124332,\n        \"num_unique_values\": 3302,\n        \"samples\": [\n          799.6624282862263,\n          3511.5915940466757\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"num_items\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1.58171040788726,\n        \"min\": 1.0,\n        \"max\": 9.0,\n        \"num_unique_values\": 9,\n        \"samples\": [\n          5.0,\n          3.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 8
        }
      ],
      "source": [
        "df1.head()"
      ],
      "id": "4LVIemy8-MRr"
    },
    {
      "cell_type": "code",
      "source": [
        "# Если мы совершали покупку (есть две записи cart-add и cart-paid с одинаковым cart_id), то мы считаем это в суммарную выручку, иначе не считаем"
      ],
      "metadata": {
        "id": "2XS4G4b8Tl37"
      },
      "id": "2XS4G4b8Tl37",
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df1_test = df1[df1['group']=='test']\n",
        "df1_control = df1[df1['group']=='control']\n",
        "\n",
        "import pandas as pd\n",
        "\n",
        "def find_check_pairs(data):\n",
        "    # Отфильтруем данные, чтобы оставить только нужные события\n",
        "    filtered_data = data[(data['event'] == 'cart-add') | (data['event'] == 'cart-paid')]\n",
        "\n",
        "    # Создадим DataFrame для хранения пар наблюдений\n",
        "    pairs_df = pd.DataFrame(columns=['user_id', 'group', 'cart_id', 'cart_paid_id', 'revenue'])\n",
        "\n",
        "    # Пройдемся по каждому пользователю\n",
        "    for user_id in filtered_data['user_id'].unique():\n",
        "        user_data = filtered_data[filtered_data['user_id'] == user_id]\n",
        "        # Пройдемся по каждому cart_id пользователя\n",
        "        for cart_id in user_data['cart_id'].unique():\n",
        "            cart_data = user_data[user_data['cart_id'] == cart_id]\n",
        "            # Проверим, есть ли у данного cart_id события \"cart-add\" и \"cart-paid\"\n",
        "            if ('cart-add' in cart_data['event'].values) and ('cart-paid' in cart_data['event'].values):\n",
        "                cart_add_row = cart_data[cart_data['event'] == 'cart-add'].iloc[0]\n",
        "                cart_paid_row = cart_data[cart_data['event'] == 'cart-paid'].iloc[0]\n",
        "                # Добавим пару наблюдений в DataFrame\n",
        "                pairs_df = pairs_df.append({'user_id': user_id,\n",
        "                                            'group': cart_add_row['group'],\n",
        "                                            'cart_id': cart_id,\n",
        "                                            'cart_paid_id': cart_paid_row['visit_id'],\n",
        "                                            'revenue': cart_add_row['amount']},\n",
        "                                           ignore_index=True)\n",
        "\n",
        "    return pairs_df\n",
        "\n",
        "\n",
        "rev_orders_test = find_check_pairs(df1_test)\n",
        "rev_orders_control = find_check_pairs(df1_control)"
      ],
      "metadata": {
        "id": "-6whqQ79Un1d"
      },
      "id": "-6whqQ79Un1d",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "rev_orders_test.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "gbqgA0hOqE-M",
        "outputId": "eeec7b12-1005-4b33-8a26-536f4060f862"
      },
      "id": "gbqgA0hOqE-M",
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "  user_id group     cart_id          cart_paid_id      revenue\n",
              "0  6W79RT  test  zd6s005p6q  84k1tk4iugavhadkbai7   673.411454\n",
              "1  6W79RT  test  keqkyd3ay9  sunaisr38ktqvulgs6ns  5055.935244\n",
              "2  6ZODD2  test  kl2m6tttzu  qg3pn8x6528gnj0fssli  2306.844187\n",
              "3  8325MF  test  pdp3irhzbb  s8m4p232f906n6m7gxe7   698.153135\n",
              "4  UVX1O1  test  ikhjkog2iy  sxip6yhdqwupf9j9bpb2   507.701473"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-cd420bbd-5ace-45cb-b80f-0ac580b8dffe\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user_id</th>\n",
              "      <th>group</th>\n",
              "      <th>cart_id</th>\n",
              "      <th>cart_paid_id</th>\n",
              "      <th>revenue</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>6W79RT</td>\n",
              "      <td>test</td>\n",
              "      <td>zd6s005p6q</td>\n",
              "      <td>84k1tk4iugavhadkbai7</td>\n",
              "      <td>673.411454</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>6W79RT</td>\n",
              "      <td>test</td>\n",
              "      <td>keqkyd3ay9</td>\n",
              "      <td>sunaisr38ktqvulgs6ns</td>\n",
              "      <td>5055.935244</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>6ZODD2</td>\n",
              "      <td>test</td>\n",
              "      <td>kl2m6tttzu</td>\n",
              "      <td>qg3pn8x6528gnj0fssli</td>\n",
              "      <td>2306.844187</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>8325MF</td>\n",
              "      <td>test</td>\n",
              "      <td>pdp3irhzbb</td>\n",
              "      <td>s8m4p232f906n6m7gxe7</td>\n",
              "      <td>698.153135</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>UVX1O1</td>\n",
              "      <td>test</td>\n",
              "      <td>ikhjkog2iy</td>\n",
              "      <td>sxip6yhdqwupf9j9bpb2</td>\n",
              "      <td>507.701473</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-cd420bbd-5ace-45cb-b80f-0ac580b8dffe')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-cd420bbd-5ace-45cb-b80f-0ac580b8dffe button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-cd420bbd-5ace-45cb-b80f-0ac580b8dffe');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-910b910d-6afa-4caf-a21c-4bf3ee63e9fd\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-910b910d-6afa-4caf-a21c-4bf3ee63e9fd')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-910b910d-6afa-4caf-a21c-4bf3ee63e9fd button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "rev_orders_test",
              "summary": "{\n  \"name\": \"rev_orders_test\",\n  \"rows\": 1000,\n  \"fields\": [\n    {\n      \"column\": \"user_id\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 959,\n        \"samples\": [\n          \"NOV4NY\",\n          \"MNP06C\",\n          \"DAYO9K\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"group\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 1,\n        \"samples\": [\n          \"test\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"cart_id\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 1000,\n        \"samples\": [\n          \"nslv1csuo9\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"cart_paid_id\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 1000,\n        \"samples\": [\n          \"tj6x17vdcjc1omjio76k\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"revenue\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 3678.4283367725957,\n        \"min\": 33.982708940048425,\n        \"max\": 55109.90049124332,\n        \"num_unique_values\": 1000,\n        \"samples\": [\n          2083.353957307967\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Мы получили таблицы для тестовой и контрольной группы, в этих таблицах гарантированно есть пользователи, у которых есть хотя бы одна запись c меткой cart-paid. Теперь добавим к этим пользователям пользователей с нулевой выручкой"
      ],
      "metadata": {
        "id": "IqMo11E4r5Q9"
      },
      "id": "IqMo11E4r5Q9"
    },
    {
      "cell_type": "code",
      "source": [
        "df1_test_total = pd.concat([rev_orders_test, df1_test[~df1_test['user_id'].isin(rev_orders_test['user_id'])]])[['user_id', 'group', 'cart_id', 'cart_paid_id', 'revenue']]\n",
        "df1_control_total = pd.concat([rev_orders_control, df1_control[~df1_control['user_id'].isin(rev_orders_control['user_id'])]])[['user_id', 'group', 'cart_id', 'cart_paid_id', 'revenue']]"
      ],
      "metadata": {
        "id": "PXfEgT0fuv-F"
      },
      "id": "PXfEgT0fuv-F",
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df1_test_total.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "Mx9k0giHv-2n",
        "outputId": "c5039611-eb28-425c-aca1-79e0b93ef83d"
      },
      "id": "Mx9k0giHv-2n",
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "  user_id group     cart_id          cart_paid_id      revenue\n",
              "0  6W79RT  test  zd6s005p6q  84k1tk4iugavhadkbai7   673.411454\n",
              "1  6W79RT  test  keqkyd3ay9  sunaisr38ktqvulgs6ns  5055.935244\n",
              "2  6ZODD2  test  kl2m6tttzu  qg3pn8x6528gnj0fssli  2306.844187\n",
              "3  8325MF  test  pdp3irhzbb  s8m4p232f906n6m7gxe7   698.153135\n",
              "4  UVX1O1  test  ikhjkog2iy  sxip6yhdqwupf9j9bpb2   507.701473"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-adeaac05-b18d-4099-aad5-89aad24e88f8\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user_id</th>\n",
              "      <th>group</th>\n",
              "      <th>cart_id</th>\n",
              "      <th>cart_paid_id</th>\n",
              "      <th>revenue</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>6W79RT</td>\n",
              "      <td>test</td>\n",
              "      <td>zd6s005p6q</td>\n",
              "      <td>84k1tk4iugavhadkbai7</td>\n",
              "      <td>673.411454</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>6W79RT</td>\n",
              "      <td>test</td>\n",
              "      <td>keqkyd3ay9</td>\n",
              "      <td>sunaisr38ktqvulgs6ns</td>\n",
              "      <td>5055.935244</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>6ZODD2</td>\n",
              "      <td>test</td>\n",
              "      <td>kl2m6tttzu</td>\n",
              "      <td>qg3pn8x6528gnj0fssli</td>\n",
              "      <td>2306.844187</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>8325MF</td>\n",
              "      <td>test</td>\n",
              "      <td>pdp3irhzbb</td>\n",
              "      <td>s8m4p232f906n6m7gxe7</td>\n",
              "      <td>698.153135</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>UVX1O1</td>\n",
              "      <td>test</td>\n",
              "      <td>ikhjkog2iy</td>\n",
              "      <td>sxip6yhdqwupf9j9bpb2</td>\n",
              "      <td>507.701473</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-adeaac05-b18d-4099-aad5-89aad24e88f8')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-adeaac05-b18d-4099-aad5-89aad24e88f8 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-adeaac05-b18d-4099-aad5-89aad24e88f8');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-d07e1036-bf77-41ef-b075-89410364277b\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-d07e1036-bf77-41ef-b075-89410364277b')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-d07e1036-bf77-41ef-b075-89410364277b button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "df1_test_total",
              "summary": "{\n  \"name\": \"df1_test_total\",\n  \"rows\": 9877,\n  \"fields\": [\n    {\n      \"column\": \"user_id\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 3295,\n        \"samples\": [\n          \"7IASW8\",\n          \"EO5888\",\n          \"FLGVOM\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"group\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 1,\n        \"samples\": [\n          \"test\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"cart_id\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 1062,\n        \"samples\": [\n          \"a2ohwet6jt\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"cart_paid_id\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 1000,\n        \"samples\": [\n          \"tj6x17vdcjc1omjio76k\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"revenue\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 3678.4283367725957,\n        \"min\": 33.982708940048425,\n        \"max\": 55109.90049124332,\n        \"num_unique_values\": 1000,\n        \"samples\": [\n          2083.353957307967\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "rev_test = df1_test_total.groupby('user_id')['revenue'].sum()\n",
        "rev_control = df1_control_total.groupby('user_id')['revenue'].sum()\n",
        "\n",
        "orders_test = df1_test_total.groupby('user_id')['cart_id'].count() # Посчитаем число покупок\n",
        "orders_control = df1_control_total.groupby('user_id')['cart_id'].count()"
      ],
      "metadata": {
        "id": "dG3Oa0CjyOxQ"
      },
      "id": "dG3Oa0CjyOxQ",
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "rev_test.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mKTlky2qyW5c",
        "outputId": "c32b7afd-4f14-455f-a863-b32e66b7e7bd"
      },
      "id": "mKTlky2qyW5c",
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "user_id\n",
              "009TG0       0.000000\n",
              "01ZW0R    5853.584375\n",
              "023T5M       0.000000\n",
              "029P4I       0.000000\n",
              "02CI39       0.000000\n",
              "Name: revenue, dtype: float64"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "metadata": {
        "id": "-TSJZ5Au4yxD",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a28bd877-23a4-42a0-c041-5fce44cae962"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Значение 70% квантиля изменилось на: 0.00\n",
            "95.0% доверительный интервал: (0.00, 137.38)\n",
            "Отличия статистически значимые: False\n"
          ]
        }
      ],
      "source": [
        "B = 10000\n",
        "alpha = 0.05\n",
        "\n",
        "def get_percentile_ci(bootstrap_stats, pe, alpha):\n",
        "    \"\"\"Строит перцентильный доверительный интервал.\"\"\"\n",
        "    left, right = np.quantile(bootstrap_stats, [alpha / 2, 1 - alpha / 2])\n",
        "    return left, right\n",
        "\n",
        "pe = np.quantile(rev_test, 0.7) - np.quantile(rev_control, 0.7)\n",
        "\n",
        "bootstrap_values_control = np.random.choice(rev_control, (B, len(rev_control)), True) # сгенерируем n примеров размера B (матрицу Bxn).\n",
        "bootstrap_metrics_control = np.quantile(bootstrap_values_control, 0.7, axis=1) # посчитаем квантиль. Получим n квантилей\n",
        "\n",
        "bootstrap_values_test = np.random.choice(rev_test, (B, len(rev_test)), True)\n",
        "bootstrap_metrics_test = np.quantile(bootstrap_values_test, 0.7, axis=1)\n",
        "\n",
        "bootstrap_stats = bootstrap_metrics_test - bootstrap_metrics_control\n",
        "ci = get_percentile_ci(bootstrap_stats, pe, alpha)\n",
        "has_effect = not (ci[0] <= 0 <= ci[1])\n",
        "\n",
        "print(f'Значение 70% квантиля изменилось на: {pe:0.2f}')\n",
        "print(f'{((1 - alpha) * 100)}% доверительный интервал: ({ci[0]:0.2f}, {ci[1]:0.2f})')\n",
        "print(f'Отличия статистически значимые: {has_effect}')"
      ],
      "id": "-TSJZ5Au4yxD"
    },
    {
      "cell_type": "markdown",
      "source": [
        "Видим, что разница точечной оценки между тестовой группой с примененным эффектом и контрольной группой, которая была без рекомендательной системы, нулевая, что говорит о том, что выручка в группе с рекомендательными системами не изменилась, эффекта нет.\n",
        "\n",
        "Мы  также видим, что доверительный интервал включает ноль, поэтому можно говорить, что различия статистически незначимы на уровне $\\alpha$. Нет смысла добавлять рекомендательные системы."
      ],
      "metadata": {
        "id": "JLlXQv2cNrgH"
      },
      "id": "JLlXQv2cNrgH"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "40e37bff-db78-430c-8d9c-cfb11a99ba4d"
      },
      "source": [
        "### Задание 3 (1 балл)\n",
        "\n",
        "Как проверить гипотезу о том, что средний чек в А и Б группе разный? Реализуйте этот алгоритм. Помните, что у одного пользователя может быть много чеков."
      ],
      "id": "40e37bff-db78-430c-8d9c-cfb11a99ba4d"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RBWYMUPLIMcS"
      },
      "source": [
        "##### Для каждого юзера для каждой группы рассчитаем средний чек, чтобы пользователи не были зависимыми. В итоге у нас должна получиться одна метрика на одного пользователя. Поэтому т-тест стьюдента будет применим в таких условиях. Проведем тест на значимость разницы средних чеков в группе А и B."
      ],
      "id": "RBWYMUPLIMcS"
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "id": "bSzs8vWSLbQ2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "07c0a398-7e2b-4edf-9324-03a8e8dd3567"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "  user_id         dt          event group              visit_id cart_id  \\\n",
              "0  3Y6FR5 2024-01-17  first-visited   NaN                   NaN     NaN   \n",
              "1  6W79RT 2024-01-17  first-visited   NaN                   NaN     NaN   \n",
              "2  6W79RT 2024-01-17     registered  test                   NaN     NaN   \n",
              "3  6W79RT 2024-01-17     shop-visit  test  3qzhu3sar2vee6rzssyr     NaN   \n",
              "4  6W79RT 2024-01-18     shop-visit  test  siglppuqke79ts5a5r4t     NaN   \n",
              "\n",
              "   amount  num_items  \n",
              "0     NaN        NaN  \n",
              "1     NaN        NaN  \n",
              "2     NaN        NaN  \n",
              "3     NaN        NaN  \n",
              "4     NaN        NaN  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-efc2dee7-2661-4a84-8d23-3bbb456db44e\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user_id</th>\n",
              "      <th>dt</th>\n",
              "      <th>event</th>\n",
              "      <th>group</th>\n",
              "      <th>visit_id</th>\n",
              "      <th>cart_id</th>\n",
              "      <th>amount</th>\n",
              "      <th>num_items</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>3Y6FR5</td>\n",
              "      <td>2024-01-17</td>\n",
              "      <td>first-visited</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>6W79RT</td>\n",
              "      <td>2024-01-17</td>\n",
              "      <td>first-visited</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>6W79RT</td>\n",
              "      <td>2024-01-17</td>\n",
              "      <td>registered</td>\n",
              "      <td>test</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>6W79RT</td>\n",
              "      <td>2024-01-17</td>\n",
              "      <td>shop-visit</td>\n",
              "      <td>test</td>\n",
              "      <td>3qzhu3sar2vee6rzssyr</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>6W79RT</td>\n",
              "      <td>2024-01-18</td>\n",
              "      <td>shop-visit</td>\n",
              "      <td>test</td>\n",
              "      <td>siglppuqke79ts5a5r4t</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-efc2dee7-2661-4a84-8d23-3bbb456db44e')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-efc2dee7-2661-4a84-8d23-3bbb456db44e button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-efc2dee7-2661-4a84-8d23-3bbb456db44e');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-4a7134c4-9f61-4cf5-a555-be7df59d6563\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-4a7134c4-9f61-4cf5-a555-be7df59d6563')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-4a7134c4-9f61-4cf5-a555-be7df59d6563 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "df1",
              "summary": "{\n  \"name\": \"df1\",\n  \"rows\": 51382,\n  \"fields\": [\n    {\n      \"column\": \"user_id\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 9054,\n        \"samples\": [\n          \"3Q129M\",\n          \"BHMY36\",\n          \"KUZOTE\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"dt\",\n      \"properties\": {\n        \"dtype\": \"date\",\n        \"min\": \"2024-01-10 00:00:00\",\n        \"max\": \"2024-01-24 00:00:00\",\n        \"num_unique_values\": 15,\n        \"samples\": [\n          \"2024-01-14 00:00:00\",\n          \"2024-01-15 00:00:00\",\n          \"2024-01-17 00:00:00\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"event\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 5,\n        \"samples\": [\n          \"registered\",\n          \"cart-paid\",\n          \"shop-visit\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"group\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 2,\n        \"samples\": [\n          \"control\",\n          \"test\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"visit_id\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 29998,\n        \"samples\": [\n          \"tsha83gw00n4k97nb0la\",\n          \"t012s0ahengc4or9smj4\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"cart_id\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 2757,\n        \"samples\": [\n          \"os3dmm6spj\",\n          \"aoj7bika3y\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"amount\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 3958.4343152104493,\n        \"min\": 26.63452376890258,\n        \"max\": 55109.90049124332,\n        \"num_unique_values\": 3302,\n        \"samples\": [\n          799.6624282862263,\n          3511.5915940466757\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"num_items\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1.58171040788726,\n        \"min\": 1.0,\n        \"max\": 9.0,\n        \"num_unique_values\": 9,\n        \"samples\": [\n          5.0,\n          3.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 17
        }
      ],
      "source": [
        "df1.head()"
      ],
      "id": "bSzs8vWSLbQ2"
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "metadata": {
        "id": "hLV82WoyPB6T"
      },
      "outputs": [],
      "source": [
        "orders_test = df1_test_total.groupby('user_id')['cart_id'].count() # Посчитаем число покупок\n",
        "orders_control = df1_control_total.groupby('user_id')['cart_id'].count()"
      ],
      "id": "hLV82WoyPB6T"
    },
    {
      "cell_type": "code",
      "source": [
        "orders_test"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "H8AAOhXsbMMF",
        "outputId": "58bc4c2a-b6cf-4bd5-ef62-80ac5d4107e3"
      },
      "id": "H8AAOhXsbMMF",
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "user_id\n",
              "009TG0    0\n",
              "01ZW0R    1\n",
              "023T5M    0\n",
              "029P4I    0\n",
              "02CI39    0\n",
              "         ..\n",
              "ZYS6GU    0\n",
              "ZYWANO    0\n",
              "ZZ55D1    1\n",
              "ZZEVTV    0\n",
              "ZZOQID    1\n",
              "Name: cart_id, Length: 3295, dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "metadata": {
        "id": "Jo5Z_dgxLg9S"
      },
      "outputs": [],
      "source": [
        "rev_test = df1_test_total.groupby('user_id')['revenue'].sum()\n",
        "rev_control = df1_control_total.groupby('user_id')['revenue'].sum()"
      ],
      "id": "Jo5Z_dgxLg9S"
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "metadata": {
        "id": "zAALk8m1NsCy"
      },
      "outputs": [],
      "source": [
        "df_merged_test = pd.merge(rev_test, orders_test, on='user_id', how='outer')\n",
        "df_merged_control = pd.merge(rev_control, orders_control, on='user_id', how='outer')\n",
        "\n",
        "# # Рассчитайте средний чек для каждого пользователя\n",
        "df_merged_test['average_order_value'] = df_merged_test['revenue'] / df_merged_test['cart_id']\n",
        "df_merged_test = df_merged_test.fillna(0, inplace=False)\n",
        "\n",
        "df_merged_control['average_order_value'] = df_merged_control['revenue'] / df_merged_control['cart_id']\n",
        "df_merged_control = df_merged_control.fillna(0, inplace=False)"
      ],
      "id": "zAALk8m1NsCy"
    },
    {
      "cell_type": "code",
      "source": [
        "df_merged_test.loc[\"6W79RT\"]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "S2nSXkZi6wTS",
        "outputId": "cceb8c9c-9c95-4ed6-9101-d4dc2fb16122"
      },
      "id": "S2nSXkZi6wTS",
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "revenue                5729.346699\n",
              "cart_id                   2.000000\n",
              "average_order_value    2864.673349\n",
              "Name: 6W79RT, dtype: float64"
            ]
          },
          "metadata": {},
          "execution_count": 22
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "average_order_test = df_merged_test['average_order_value']\n",
        "average_order_control = df_merged_control['average_order_value']"
      ],
      "metadata": {
        "id": "GMYDJCB-o0gU"
      },
      "id": "GMYDJCB-o0gU",
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.hist(average_order_test, bins=20, alpha=0.5, label='Тестовая группа')\n",
        "plt.hist(average_order_control, bins=20, alpha=0.5, label='Контрольная группа')\n",
        "plt.xlabel('Значение среднего чека')\n",
        "plt.ylabel('Частота')\n",
        "plt.title('Распределение значений среднего чека для тестовой и контрольной групп')\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "id": "baF9kcY2zdRO",
        "outputId": "e3c82cf5-6066-41fa-909c-59a73f7405be"
      },
      "id": "baF9kcY2zdRO",
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Видим, что распределение средних чеков в тестовой и контрольной группах отличаются, поэтому не будем применять среднее средних, важно семплировать данные по объектам в таком случае."
      ],
      "metadata": {
        "id": "ImNPB7tH0kux"
      },
      "id": "ImNPB7tH0kux"
    },
    {
      "cell_type": "code",
      "source": [
        "df_merged_test = pd.merge(rev_test, orders_test, on='user_id', how='outer')\n",
        "df_merged_control = pd.merge(rev_control, orders_control, on='user_id', how='outer')"
      ],
      "metadata": {
        "id": "pScxHqBrFMRR"
      },
      "id": "pScxHqBrFMRR",
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Мы нашли сумму заказов и число чеков по пользователю для тестовой и контрольной выборок, теперь давайте проведем тест с помощью бутстрапа: будем находить значение статистики (разницы между средними чеками пользователей двух групп), строить доверительный интервал и смотреть, принадлежит ли 0 данному интервалу. Если не принадлежит, значит статистика значимо отличается от нуля, значит есть значимые различия между средним чеком в двух группах (на 5% уровне значимости)."
      ],
      "metadata": {
        "id": "_oPQDL87i3v_"
      },
      "id": "_oPQDL87i3v_"
    },
    {
      "cell_type": "code",
      "source": [
        "def get_percentile_ci(bootstrap_stats, alpha):\n",
        "    \"\"\"Строит перцентильный доверительный интервал.\"\"\"\n",
        "    left, right = np.quantile(bootstrap_stats, [alpha / 2, 1 - alpha / 2])\n",
        "    return left, right\n",
        "\n",
        "def check_bootstrap(a, b, n_iter, alpha):\n",
        "    \"\"\"Оценивает значимость отличий с помощью бутстрепа.\n",
        "\n",
        "    Если отличия значимые, то возвращает 1, иначе 0.\n",
        "    \"\"\"\n",
        "    # вычисляем стоимость и количество покупок клиентов\n",
        "    # xy_a = np.array([[sum(values), len(values)] for values in a]) # берется список с данными о покупках, затем переделывается в массив размера len(a) строк, 2 столбца (сумма и количество заказов)\n",
        "    # xy_b = np.array([[sum(values), len(values)] for values in b])\n",
        "    # генерируем случайные индексы для выбора подмножеств данных\n",
        "    len_a = len(a) # sample_size\n",
        "    len_b = len(b)\n",
        "    indexes_a = np.random.choice(\n",
        "        np.arange(len_a), size=(n_iter, len_a), replace=True # Генерим n_iter*len_a сэмплов - n_iter рандомных последовательностей индексов\n",
        "    )\n",
        "    indexes_b = np.random.choice(\n",
        "        np.arange(len_b), size=(n_iter, len_b), replace=True\n",
        "    )\n",
        "    bootstrap_stats = []\n",
        "    for idx_a, idx_b in zip(indexes_a, indexes_b): # idx_a это одно число\n",
        "        bootstrap_xy_a = a.iloc[idx_a] # берем n пользователей из группы А и B\n",
        "        bootstrap_xy_b = b.iloc[idx_b]\n",
        "        # считаем разницу метрик отношения\n",
        "        bootstrap_stat = (\n",
        "            bootstrap_xy_b['revenue'].sum() / bootstrap_xy_b['cart_id'].sum() # считаем разницу в среднем чеке по всем пользователям между группами и повторяем indexes_a.shape[1] раз, то есть n_iter раз считаем разницу в средних чеках между всеми пользователями и добавляем статистику в bootstrap_stats\n",
        "            - bootstrap_xy_a['revenue'].sum() / bootstrap_xy_a['cart_id'].sum()\n",
        "        )\n",
        "        bootstrap_stats.append(bootstrap_stat)\n",
        "    # строим доверительный интервал и оцениваем значимость отличий\n",
        "    ci = get_percentile_ci(bootstrap_stats, alpha)\n",
        "    has_effect = 1 - (ci[0] < 0 < ci[1])\n",
        "    return has_effect\n",
        "\n",
        "alpha = 0.05               # допустимая вероятность ошибки I рода\n",
        "n_iter = 1000             # количество итераций бутстрепа\n",
        "\n",
        "effects = []\n",
        "for _ in range(100):\n",
        "  has_effect = check_bootstrap(df_merged_control, df_merged_test, n_iter, alpha)\n",
        "  effects.append(has_effect)\n",
        "\n",
        "error_rate = np.mean(np.array(effects) == 1)\n",
        "print(f'Лежит ли полученная статистика разницы в доверительном интервале: {not bool(error_rate)}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xET65nnzlotb",
        "outputId": "b6875c73-3df0-431e-ff3e-1819a452fae2"
      },
      "id": "xET65nnzlotb",
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Лежит ли полученная статистика разницы в доверительном интервале: True\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Мы видим, что при большом числе проведенных бутстрапированных тестов, наша бутстрапированная статистика не лежит в доверительном интервале, это значит, что наша статистика значимо отличается от нуля"
      ],
      "metadata": {
        "id": "li-5dm4FT8a8"
      },
      "id": "li-5dm4FT8a8"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cb810f77-ebd9-439f-9ce6-c56babf7d02a"
      },
      "source": [
        "### Задание 4 (1 балл)\n",
        "\n",
        "Мы заменили все фотографии мыла на сайте более привлекательными снимками от профессионального фотографа и отфотошопили их. Результаты теста в файле `ctr_test_logs.pkl`. Как такое изменение повлияло на конверсию из визита в добавление в корзину? Проведите соответствующий тест и не забудьте о том, что имеете дело с метрикой отношения.\n",
        "\n",
        "**P.s. в этом задании решение методом бутстрапа не будет засчитываться**"
      ],
      "id": "cb810f77-ebd9-439f-9ce6-c56babf7d02a"
    },
    {
      "cell_type": "code",
      "source": [
        "# Прямая ссылка на скачивание файла\n",
        "url = 'https://drive.google.com/uc?id=1uWsUDNinvxx0UKn5CRLgB5vD0oUwz_Qe'\n",
        "\n",
        "# Загрузка файла\n",
        "response = requests.get(url)\n",
        "\n",
        "# Чтение данных pickle\n",
        "df2 = pickle.loads(response.content)"
      ],
      "metadata": {
        "id": "MIGxF2a5VJLG"
      },
      "id": "MIGxF2a5VJLG",
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df2.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "UuEjT29hXVfa",
        "outputId": "f780d767-7804-4675-a158-87498b0e9c27"
      },
      "id": "UuEjT29hXVfa",
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "  user_id         dt          event group              visit_id cart_id  \\\n",
              "0  Q9IK2A 2024-02-25  first-visited   NaN                   NaN     NaN   \n",
              "1  Q9IK2A 2024-02-25     registered     b                   NaN     NaN   \n",
              "2  Q9IK2A 2024-02-26     shop-visit     b  k4yi0jd4clxduvmsk04g     NaN   \n",
              "3  XVXL0K 2024-02-13  first-visited   NaN                   NaN     NaN   \n",
              "4  XVXL0K 2024-02-14     shop-visit     a  sp3hls8m82xwudv1qgms     NaN   \n",
              "\n",
              "   amount  num_items  \n",
              "0     NaN        NaN  \n",
              "1     NaN        NaN  \n",
              "2     NaN        NaN  \n",
              "3     NaN        NaN  \n",
              "4     NaN        NaN  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-e3ed50aa-179b-4adf-8f9d-04efb5f25a55\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user_id</th>\n",
              "      <th>dt</th>\n",
              "      <th>event</th>\n",
              "      <th>group</th>\n",
              "      <th>visit_id</th>\n",
              "      <th>cart_id</th>\n",
              "      <th>amount</th>\n",
              "      <th>num_items</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Q9IK2A</td>\n",
              "      <td>2024-02-25</td>\n",
              "      <td>first-visited</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Q9IK2A</td>\n",
              "      <td>2024-02-25</td>\n",
              "      <td>registered</td>\n",
              "      <td>b</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Q9IK2A</td>\n",
              "      <td>2024-02-26</td>\n",
              "      <td>shop-visit</td>\n",
              "      <td>b</td>\n",
              "      <td>k4yi0jd4clxduvmsk04g</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>XVXL0K</td>\n",
              "      <td>2024-02-13</td>\n",
              "      <td>first-visited</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>XVXL0K</td>\n",
              "      <td>2024-02-14</td>\n",
              "      <td>shop-visit</td>\n",
              "      <td>a</td>\n",
              "      <td>sp3hls8m82xwudv1qgms</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-e3ed50aa-179b-4adf-8f9d-04efb5f25a55')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-e3ed50aa-179b-4adf-8f9d-04efb5f25a55 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-e3ed50aa-179b-4adf-8f9d-04efb5f25a55');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-b7b039bd-aac0-4067-8c5a-12fead3f260b\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-b7b039bd-aac0-4067-8c5a-12fead3f260b')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-b7b039bd-aac0-4067-8c5a-12fead3f260b button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "df2",
              "summary": "{\n  \"name\": \"df2\",\n  \"rows\": 29115,\n  \"fields\": [\n    {\n      \"column\": \"user_id\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 4766,\n        \"samples\": [\n          \"HOB7HP\",\n          \"41VD1Z\",\n          \"XY9ISJ\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"dt\",\n      \"properties\": {\n        \"dtype\": \"date\",\n        \"min\": \"2024-02-10 00:00:00\",\n        \"max\": \"2024-02-28 00:00:00\",\n        \"num_unique_values\": 19,\n        \"samples\": [\n          \"2024-02-25 00:00:00\",\n          \"2024-02-22 00:00:00\",\n          \"2024-02-27 00:00:00\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"event\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 5,\n        \"samples\": [\n          \"registered\",\n          \"cart-paid\",\n          \"shop-visit\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"group\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 2,\n        \"samples\": [\n          \"a\",\n          \"b\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"visit_id\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 18032,\n        \"samples\": [\n          \"5m9d7klhs8wtrwccm6a5\",\n          \"9nilvu4fv4zzr76audzn\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"cart_id\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 1494,\n        \"samples\": [\n          \"cxivcxlmpa\",\n          \"bdcn7pnx5n\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"amount\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 4865.50664975718,\n        \"min\": 40.50236964295237,\n        \"max\": 62808.28483559689,\n        \"num_unique_values\": 1850,\n        \"samples\": [\n          1571.2279894732726,\n          20731.13946846457\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"num_items\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1.6925591292318973,\n        \"min\": 1.0,\n        \"max\": 9.0,\n        \"num_unique_values\": 9,\n        \"samples\": [\n          5.0,\n          2.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Метрика конверсии как метрика отношения может привести к зависимости данных, поэтому обычный тест Стьюдента не применим. Применим дельта метод. Причем обратим внимание, что $X_i$ зависимы между собой, потому что у одного и того же пользователя может быть несколько покупок."
      ],
      "metadata": {
        "id": "1zAg3t-lYToc"
      },
      "id": "1zAg3t-lYToc"
    },
    {
      "cell_type": "markdown",
      "source": [
        "![image.png](data:image/png;base64,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)"
      ],
      "metadata": {
        "id": "ufyp5MAfgz8w"
      },
      "id": "ufyp5MAfgz8w"
    },
    {
      "cell_type": "markdown",
      "source": [
        "$t=\\frac{\\mathcal{R}_B-\\mathcal{R}_A}{\\sqrt{\\mathbb{V}\\left(\\mathcal{R}_A\\right)+\\mathbb{V}\\left(\\mathcal{R}_B\\right)}} \\xrightarrow{N \\rightarrow \\inf } \\mathcal{N}(0,1)$"
      ],
      "metadata": {
        "id": "0Zfy0P-Hg5gY"
      },
      "id": "0Zfy0P-Hg5gY"
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Перед ковариацией у нас появится тоже $\\frac{1}{N}$, потому что визит на сайт и добавление товара в корзину являются **зависимыми случайными величинами**: будем предполагать, что чем больше человек посещал сайт, тем более он \"прогретый\" и тем больше вероятность того, что он купит товар (также если он купил товар = посетил сайт перед покупкой, тогда случайные величины 100% зависимые, без всяких предположений)."
      ],
      "metadata": {
        "id": "0BS4x_fKk3hM"
      },
      "id": "0BS4x_fKk3hM"
    },
    {
      "cell_type": "code",
      "source": [
        "# Фильтрация данных для группы A\n",
        "group_a = df2[df2['group'] == 'a']\n",
        "\n",
        "# Фильтрация данных для группы B\n",
        "group_b = df2[df2['group'] == 'b']\n",
        "\n",
        "# Подсчет числа визитов для каждого пользователя в группе A\n",
        "visits_a = group_a[group_a['event'] == 'shop-visit'].groupby('user_id').size().reset_index(name='visits') # Подсчет числа визитов\n",
        "cart_additions_a = group_a[group_a['event'] == 'cart-add'].groupby('user_id').size().reset_index(name='cart_additions') # Подсчет числа добавлений в корзину\n",
        "df_a = pd.merge(visits_a, cart_additions_a, on='user_id', how='outer')\n",
        "df_a['visits'] = df_a['visits'].fillna(0)\n",
        "df_a['cart_additions'] = df_a['cart_additions'].fillna(0)\n",
        "df_a['visits'] += df_a['cart_additions']\n",
        "\n",
        "# Подсчет числа визитов для каждого пользователя в группе B\n",
        "visits_b = group_b[group_b['event'] == 'shop-visit'].groupby('user_id').size().reset_index(name='visits') # Подсчет числа визитов\n",
        "cart_additions_b = group_b[group_b['event'] == 'cart-add'].groupby('user_id').size().reset_index(name='cart_additions') # Подсчет числа добавлений в корзину\n",
        "df_b = pd.merge(visits_b, cart_additions_b, on='user_id', how='outer')\n",
        "df_b['visits'] = df_b['visits'].fillna(0)\n",
        "df_b['cart_additions'] = df_b['cart_additions'].fillna(0)\n",
        "df_b['visits'] += df_b['cart_additions']"
      ],
      "metadata": {
        "id": "YWbEs4rorv4b"
      },
      "id": "YWbEs4rorv4b",
      "execution_count": 29,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_a.loc[df_a['user_id'] == 'NIFXUJ']"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 81
        },
        "id": "acFPCS9CtMxV",
        "outputId": "ecf5890d-aec9-4f84-8391-9446867bd597"
      },
      "id": "acFPCS9CtMxV",
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "    user_id  visits  cart_additions\n",
              "976  NIFXUJ     2.0             1.0"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-d6a1837e-2f6a-44e7-b43b-59c343c460a0\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user_id</th>\n",
              "      <th>visits</th>\n",
              "      <th>cart_additions</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>976</th>\n",
              "      <td>NIFXUJ</td>\n",
              "      <td>2.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-d6a1837e-2f6a-44e7-b43b-59c343c460a0')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-d6a1837e-2f6a-44e7-b43b-59c343c460a0 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-d6a1837e-2f6a-44e7-b43b-59c343c460a0');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "summary": "{\n  \"name\": \"df_a\",\n  \"rows\": 1,\n  \"fields\": [\n    {\n      \"column\": \"user_id\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 1,\n        \"samples\": [\n          \"NIFXUJ\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"visits\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": null,\n        \"min\": 2.0,\n        \"max\": 2.0,\n        \"num_unique_values\": 1,\n        \"samples\": [\n          2.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"cart_additions\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": null,\n        \"min\": 1.0,\n        \"max\": 1.0,\n        \"num_unique_values\": 1,\n        \"samples\": [\n          1.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from scipy import stats\n",
        "\n",
        "def check_delta_method(a, b):\n",
        "    \"\"\"Проверка гипотезы с помощью дельта-метода.\n",
        "\n",
        "    Возвращает p-value.\n",
        "    \"\"\"\n",
        "    # ИЗ ВИЗИТА В ДОБАВЛЕНИЕ В КОРЗИНУ\n",
        "    dict_ = {}\n",
        "    for data, group_name in [(a, 'a',), (b, 'b',)]:\n",
        "        n_user = len(data)  # Длина массива а\n",
        "        array_x = np.array(data[\"cart_additions\"]) # Число добавлений в корзину для каждого пользователя\n",
        "        array_y = np.array(data[\"visits\"]) # Визиты для каждого пользователя\n",
        "        mean_x, mean_y = np.mean(array_x), np.mean(array_y) # Считаем среднее\n",
        "        var_x, var_y = np.var(array_x), np.var(array_y) # Считаем дисперсию\n",
        "        cov_xy = np.cov(array_x, array_y)[0, 1] # считаем ковариацию между X и Y\n",
        "        # точечная оценка метрики\n",
        "        pe_metric = np.sum(array_x) / np.sum(array_y)\n",
        "        # оценка дисперсии метрики\n",
        "        var_metric = (\n",
        "            var_x / mean_y ** 2\n",
        "            - 2 * (mean_x/ mean_y ** 3) * 1/n_user * cov_xy\n",
        "            + (mean_x ** 2 / mean_y ** 4) * var_y\n",
        "        ) / n_user\n",
        "        dict_[f'pe_metric_{group_name}'] = pe_metric\n",
        "        dict_[f'var_metric_{group_name}'] = var_metric\n",
        "    var = dict_['var_metric_a'] + dict_['var_metric_b']\n",
        "    delta = dict_['pe_metric_b'] - dict_['pe_metric_a']\n",
        "    t = delta / np.sqrt(var)\n",
        "    pvalue = (1 - stats.norm.cdf(np.abs(t))) * 2\n",
        "    return pvalue\n",
        "\n",
        "alpha = 0.05           # допустимая вероятность ошибки I рода\n",
        "\n",
        "pvalue = check_delta_method(df_a, df_b)\n",
        "print(f'p-value теста на равенство метрик отношения: {pvalue}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DqlKhTyBhLl1",
        "outputId": "887ecde5-23ae-4ef9-a0e1-ebe489e01da7"
      },
      "id": "DqlKhTyBhLl1",
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "p-value теста на равенство метрик отношения: 0.002133845712948945\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Мы считаем p-value и видим, что есть различие между двумя группами, на уровне значимости ошибки 5%, что говорит о том, что использование фотошопа и красивых картинок изменило конверсию пользователей."
      ],
      "metadata": {
        "id": "yBNYCpGr2Zou"
      },
      "id": "yBNYCpGr2Zou"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "26a1c8f4-4992-47ad-88e3-a85848ef205c"
      },
      "source": [
        "### Задание 5 (1 балл)"
      ],
      "id": "26a1c8f4-4992-47ad-88e3-a85848ef205c"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "f2537532-30a1-42dc-af89-6c1843bf67ad"
      },
      "source": [
        "Придумайте пример, в котором тест Манна-Уитни отвергает нулевую гипотезу о равенстве для распределений с равными медианами.\n",
        "\n",
        "Постройте распределение p-value и изучите долю ошибок первого рода. Объясните, почему получился такой результат. Что тогда проверяет тест Манна-Уитни?"
      ],
      "id": "f2537532-30a1-42dc-af89-6c1843bf67ad"
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "from scipy.stats import mannwhitneyu\n",
        "\n",
        "# Генерируем выборки\n",
        "np.random.seed(42)\n",
        "sample_size = 1000\n",
        "\n",
        "sample1 = stats.uniform(loc=-1, scale=2).rvs(sample_size)\n",
        "sample2 = stats.uniform(loc=-50, scale=100).rvs(sample_size)  # draw 1000 samples\n",
        "\n",
        "# Выполняем тест Манна-Уитни\n",
        "statistic, p_value = mannwhitneyu(sample1, sample2, alternative='two-sided')\n",
        "\n",
        "# Выводим результаты\n",
        "print(f\"Статистика теста Манна-Уитни: {statistic}\")\n",
        "print(f\"p-value: {p_value}\")\n",
        "\n",
        "# Построим гистограммы распределения p-value\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "p_values = []\n",
        "for _ in range(1000):\n",
        "    sample1 = stats.uniform(loc=-1, scale=2).rvs(sample_size)\n",
        "    sample2 = stats.uniform(loc=-100, scale=200).rvs(sample_size)  # draw 1000 samples\n",
        "    _, p_value = mannwhitneyu(sample1, sample2, alternative='two-sided')\n",
        "    p_values.append(p_value)\n",
        "\n",
        "print(f'Ошибка первого рода равна: {np.sum(np.array(p_values) < alpha)/len(p_values)}')\n",
        "\n",
        "plt.hist(p_values, bins=20)\n",
        "plt.title('Распределение p-value')\n",
        "plt.xlabel('p-value')\n",
        "plt.ylabel('Частота')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 527
        },
        "id": "qvqqG02--owe",
        "outputId": "90ee2e56-16eb-46c9-c001-76cdb82f3654"
      },
      "id": "qvqqG02--owe",
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Статистика теста Манна-Уитни: 481600.0\n",
            "p-value: 0.15419681544312153\n",
            "Ошибка первого рода равна: 0.096\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from scipy.stats import kstest, uniform\n",
        "\n",
        "def is_uniform(sample, significance_level):\n",
        "    # Perform the Kolmogorov-Smirnov test\n",
        "    _, p_value = kstest(sample, uniform.cdf)\n",
        "\n",
        "    return p_value\n",
        "\n",
        "significance_level = 0.05\n",
        "# Example usage\n",
        "p_value = is_uniform(p_values, significance_level)\n",
        "\n",
        "if p_value > significance_level:\n",
        "    print(\"Uniform\")  # Sample distribution is likely uniform\n",
        "else:\n",
        "    print(\"Not Uniform\")  # Sample distribution is not likely uniform\n",
        "\n",
        "p_value"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hfTkz48MAPdv",
        "outputId": "e03d78d1-f35e-47c3-9056-a99a1599091c"
      },
      "id": "hfTkz48MAPdv",
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Not Uniform\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "7.111003109332467e-06"
            ]
          },
          "metadata": {},
          "execution_count": 33
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Мы видим, что нулевая гипотеза отвергается на уровне значимости 5%, p-value распределен неравномерно, хотя похож на равномерное распределение, а доля ошибок первого рода больше 5% уровня значимости, что означает, что нулевая гипотеза отвергается в пользу альтернативной для этих двух распределений, хотя мы знаем, что медианы и матожидания у распределений двух равномерных распределений из проведенного теста одинаковы (Мы взяли равномерное распределение $U~[-1;1]$ и $U~[-100;100]$. Это значит, что под капотом Манна-Уитни лежит вовсе не гипотеза о равенстве медиан, а гипотеза о равенстве распределений, для этого проведем другой тест, взяв уже одинаковые распределения:"
      ],
      "metadata": {
        "id": "O7JcnPsRPcp5"
      },
      "id": "O7JcnPsRPcp5"
    },
    {
      "cell_type": "code",
      "source": [
        "# Генерируем выборки\n",
        "np.random.seed(42)\n",
        "sample_size = 1000\n",
        "\n",
        "sample1 = stats.uniform(loc=-1, scale=2).rvs(sample_size)\n",
        "sample2 = stats.uniform(loc=-1, scale=2).rvs(sample_size)  # draw 1000 samples\n",
        "\n",
        "# Выполняем тест Манна-Уитни\n",
        "statistic, p_value = mannwhitneyu(sample1, sample2, alternative='two-sided')\n",
        "\n",
        "# Выводим результаты\n",
        "print(f\"Статистика теста Манна-Уитни: {statistic}\")\n",
        "print(f\"p-value: {p_value}\")\n",
        "\n",
        "# Построим гистограммы распределения p-value\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "p_values = []\n",
        "for _ in range(1000):\n",
        "    sample1 = stats.uniform(loc=-1, scale=2).rvs(sample_size)\n",
        "    sample2 = stats.uniform(loc=-1, scale=2).rvs(sample_size)  # draw 1000 samples\n",
        "    _, p_value = mannwhitneyu(sample1, sample2, alternative='two-sided')\n",
        "    p_values.append(p_value)\n",
        "\n",
        "print(f'Ошибка первого рода равна: {np.sum(np.array(p_values) < alpha)/len(p_values)}')\n",
        "\n",
        "plt.hist(p_values, bins=20)\n",
        "plt.title('Распределение p-value')\n",
        "plt.xlabel('p-value')\n",
        "plt.ylabel('Частота')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 527
        },
        "id": "N1xW1fxzT2tr",
        "outputId": "39a5994d-92af-49d8-d04b-23e705cf6a4a"
      },
      "id": "N1xW1fxzT2tr",
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Статистика теста Манна-Уитни: 483550.0\n",
            "p-value: 0.20271503318238437\n",
            "Ошибка первого рода равна: 0.05\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Вот, теперь ошибка первого рода равна 5%, p-value больше требуемого уровня значимости, что говорит о том, что нулевая гипотеза о равенстве распределений не отвергается на 5% уровне значимости, p-value распределен равномерно. Наши предположения о том, какая нулевая гипотеза используется под капотом теста Манна-Уитни подтверждается."
      ],
      "metadata": {
        "id": "aa-UoufyUdr3"
      },
      "id": "aa-UoufyUdr3"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "077f104a"
      },
      "source": [
        "### Задание 6 (1.5 балл)\n",
        "\n",
        "Вернемся к метрике revenue per user. Вам даны две очень похожие группы (файл `aa_test_logs.pkl`).\n",
        "\n",
        "1. Постройте распределение p-value для теста на равенство средних между группами.\n",
        "2. Найдите, на каком уровне искусственно добавленного эффекта средняя выручка в группах станет различима. Ошибку первого рода считайте заданной и равной 0.05. Постройте график зависимости эффекта и ошибки 2 рода."
      ],
      "id": "077f104a"
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### То есть надо изобразить p-value на равенство средних между группами, выяснить, на каком уровне effect средняя выручка станет различима, построить зависимость $\\beta$ и $\\varepsilon$"
      ],
      "metadata": {
        "id": "PTU2J6mD7i0p"
      },
      "id": "PTU2J6mD7i0p"
    },
    {
      "cell_type": "code",
      "source": [
        "# Прямая ссылка на скачивание файла\n",
        "url = 'https://drive.google.com/uc?id=1jJSRM0LLr-U3QDH954PcTqTr4lRpyZFY'\n",
        "\n",
        "# Загрузка файла\n",
        "response = requests.get(url)\n",
        "\n",
        "# Чтение данных pickle\n",
        "df3 = pickle.loads(response.content)"
      ],
      "metadata": {
        "id": "Y6k2ITA1bAPr"
      },
      "id": "Y6k2ITA1bAPr",
      "execution_count": 35,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# df3.fillna(0, inplace=True)\n",
        "df3.head()"
      ],
      "metadata": {
        "id": "66m07DIwA8Ao",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "1d6c9bad-d927-41e8-b016-f3db2c293ecd"
      },
      "id": "66m07DIwA8Ao",
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "  user_id         dt          event   group              visit_id cart_id  \\\n",
              "0  DK452H 2023-12-23  first-visited     NaN                   NaN     NaN   \n",
              "1  ED5SC3 2023-12-19  first-visited     NaN                   NaN     NaN   \n",
              "2  ED5SC3 2023-12-29     shop-visit  second  npdt49xc6xb1fcbmlh0t     NaN   \n",
              "3  GD3FMD 2023-12-02  first-visited     NaN                   NaN     NaN   \n",
              "4  GD3FMD 2023-12-02     shop-visit  second  kqsakfrjq5ys6kb0s8d0     NaN   \n",
              "\n",
              "   amount  num_items  \n",
              "0     NaN        NaN  \n",
              "1     NaN        NaN  \n",
              "2     NaN        NaN  \n",
              "3     NaN        NaN  \n",
              "4     NaN        NaN  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-35c3711a-a1a3-4b91-b344-c7c422e4ed23\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user_id</th>\n",
              "      <th>dt</th>\n",
              "      <th>event</th>\n",
              "      <th>group</th>\n",
              "      <th>visit_id</th>\n",
              "      <th>cart_id</th>\n",
              "      <th>amount</th>\n",
              "      <th>num_items</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>DK452H</td>\n",
              "      <td>2023-12-23</td>\n",
              "      <td>first-visited</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>ED5SC3</td>\n",
              "      <td>2023-12-19</td>\n",
              "      <td>first-visited</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>ED5SC3</td>\n",
              "      <td>2023-12-29</td>\n",
              "      <td>shop-visit</td>\n",
              "      <td>second</td>\n",
              "      <td>npdt49xc6xb1fcbmlh0t</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>GD3FMD</td>\n",
              "      <td>2023-12-02</td>\n",
              "      <td>first-visited</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>GD3FMD</td>\n",
              "      <td>2023-12-02</td>\n",
              "      <td>shop-visit</td>\n",
              "      <td>second</td>\n",
              "      <td>kqsakfrjq5ys6kb0s8d0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-35c3711a-a1a3-4b91-b344-c7c422e4ed23')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-35c3711a-a1a3-4b91-b344-c7c422e4ed23 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-35c3711a-a1a3-4b91-b344-c7c422e4ed23');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-4149cc11-459f-4ac4-9a10-284d8a608cca\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-4149cc11-459f-4ac4-9a10-284d8a608cca')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-4149cc11-459f-4ac4-9a10-284d8a608cca button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "df3",
              "summary": "{\n  \"name\": \"df3\",\n  \"rows\": 24393,\n  \"fields\": [\n    {\n      \"column\": \"user_id\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 7855,\n        \"samples\": [\n          \"K4OOAC\",\n          \"7CMP77\",\n          \"O424BW\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"dt\",\n      \"properties\": {\n        \"dtype\": \"date\",\n        \"min\": \"2023-12-01 00:00:00\",\n        \"max\": \"2023-12-31 00:00:00\",\n        \"num_unique_values\": 31,\n        \"samples\": [\n          \"2023-12-05 00:00:00\",\n          \"2023-12-18 00:00:00\",\n          \"2023-12-09 00:00:00\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"event\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 5,\n        \"samples\": [\n          \"shop-visit\",\n          \"cart-paid\",\n          \"registered\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"group\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 2,\n        \"samples\": [\n          \"first\",\n          \"second\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"visit_id\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 9801,\n        \"samples\": [\n          \"ah2alrowh6rydrz1ve24\",\n          \"wvswdcnhb1bxwgkd6cuz\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"cart_id\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 106,\n        \"samples\": [\n          \"vba2ftt2ya\",\n          \"jaq5yg7wdl\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"amount\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1758.7867274426826,\n        \"min\": 81.91918914780506,\n        \"max\": 9222.711177117493,\n        \"num_unique_values\": 107,\n        \"samples\": [\n          1081.0817864861267,\n          741.5318247664223\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"num_items\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1.386192841653518,\n        \"min\": 1.0,\n        \"max\": 8.0,\n        \"num_unique_values\": 7,\n        \"samples\": [\n          3.0,\n          1.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Разобьем на тест и контроль\n",
        "df3_test = df3[df3['group']=='first']\n",
        "df3_control = df3[df3['group']=='second']\n",
        "\n",
        "# Найдем пользователей, которые добавили товары в корзину и еще оплатили их\n",
        "rev_orders_test = find_check_pairs(df3_test)\n",
        "rev_orders_control = find_check_pairs(df3_control)\n",
        "\n",
        "# Добавим челиков, которые ничего не покупали и которых нет в списке что-то купивших челиков\n",
        "df3_test_total = pd.concat([rev_orders_test, df3_test[~df3_test['user_id'].isin(rev_orders_test['user_id'])]])[['user_id', 'group', 'cart_id', 'cart_paid_id', 'revenue']]\n",
        "df3_control_total = pd.concat([rev_orders_control, df3_control[~df3_control['user_id'].isin(rev_orders_control['user_id'])]])[['user_id', 'group', 'cart_id', 'cart_paid_id', 'revenue']]"
      ],
      "metadata": {
        "id": "95qNrTvMBoli"
      },
      "id": "95qNrTvMBoli",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Посчитаем поюзерную сумму выручек\n",
        "rev_first = df3_test_total.groupby('user_id')['revenue'].sum()\n",
        "rev_second = df3_control_total.groupby('user_id')['revenue'].sum()\n",
        "rev_first = rev_first.fillna(0, inplace = False)\n",
        "rev_second = rev_second.fillna(0, inplace = False)"
      ],
      "metadata": {
        "id": "PCSvGZKYBT2T"
      },
      "id": "PCSvGZKYBT2T",
      "execution_count": 42,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Теперь мы устранили зависимость между данными, проведем теперь t-test с помощью бутстрапа, потому что теперь предпосылки для проведения этого теста выполняются."
      ],
      "metadata": {
        "id": "pN6Dc6_CDtko"
      },
      "id": "pN6Dc6_CDtko"
    },
    {
      "cell_type": "code",
      "source": [
        "def bootstrap_test(a, b, n_bootstrap=10000, sample_size=None, effect=0):\n",
        "    \"\"\"Проводим бутстрап-тестирование, возвращаем список p-value.\"\"\"\n",
        "\n",
        "    pvalues = []\n",
        "    for _ in range(n_bootstrap):\n",
        "        a_sample = np.random.choice(a, size=sample_size, replace=True)\n",
        "        b_with_effect = b + effect\n",
        "        b_sample = np.random.choice(b_with_effect, size=sample_size, replace=True)\n",
        "        t_stat, pvalue = stats.ttest_ind(a_sample, b_sample)\n",
        "        pvalues.append(pvalue)\n",
        "\n",
        "    return np.array(pvalues)\n",
        "\n",
        "# Пример использования функции\n",
        "pvalues_bootstrap = bootstrap_test(rev_first, rev_second, n_bootstrap=1000, sample_size=1000, effect=0)\n",
        "\n",
        "# Визуализация результатов\n",
        "plt.hist(pvalues_bootstrap, bins=50, color='cyan', edgecolor='black')\n",
        "plt.xlabel('P-value')\n",
        "plt.ylabel('Частота')\n",
        "plt.title('Гистограмма P-value бутстрап-тестирования')\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "id": "XJOsyEKrdL6I",
        "outputId": "3527b100-dac8-4add-e29f-e39cc3f68d78"
      },
      "id": "XJOsyEKrdL6I",
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from scipy.stats import kstest, uniform\n",
        "\n",
        "def is_uniform(sample, significance_level):\n",
        "    # Perform the Kolmogorov-Smirnov test\n",
        "    _, p_value = kstest(sample, uniform.cdf)\n",
        "\n",
        "    return p_value\n",
        "\n",
        "significance_level = 0.05\n",
        "# Example usage\n",
        "p_value = is_uniform(pvalues_bootstrap.flatten(), significance_level)\n",
        "\n",
        "if p_value > significance_level:\n",
        "    print(\"Uniform\")  # Sample distribution is likely uniform\n",
        "else:\n",
        "    print(\"Not Uniform\")  # Sample distribution is not likely uniform\n",
        "\n",
        "p_value"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NnBaBkHLGEpV",
        "outputId": "85f38d89-4c43-4c48-a7cb-09de6c5a00a1"
      },
      "id": "NnBaBkHLGEpV",
      "execution_count": 44,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Uniform\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.3581938638877128"
            ]
          },
          "metadata": {},
          "execution_count": 44
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "from scipy import stats\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "def calculate_power(a, b, effect, alpha=0.05, n_bootstrap=10000, sample_size=None):\n",
        "    \"\"\"Рассчитываем мощность теста для заданного эффекта.\"\"\"\n",
        "    if sample_size is None:\n",
        "        sample_size = min(len(a), len(b))\n",
        "\n",
        "    n_reject = 0 # счетчик для подсчета числа отклонений от нулевой гипотезы\n",
        "    for _ in range(n_bootstrap):\n",
        "        a_sample = np.random.choice(a, size=sample_size, replace=True)\n",
        "        b_with_effect = b + effect\n",
        "        b_sample = np.random.choice(b_with_effect, size=sample_size, replace=True)\n",
        "        t_stat, pvalue = stats.ttest_ind(a_sample, b_sample)\n",
        "        if pvalue < alpha: # если нулевая гипотеза отвергается в пользу альтернативной (ищем, в скольких случаях мы отвергли нулевую гипотезу о равенстве средних, при том что они реально разные - мощность считаем)\n",
        "            n_reject += 1\n",
        "\n",
        "    power = n_reject / n_bootstrap\n",
        "    return power\n",
        "\n",
        "effects = np.linspace(0, 1, 20)\n",
        "powers = [calculate_power(rev_first, rev_second, eff) for eff in effects]\n",
        "beta = 1 - np.array(powers)\n",
        "\n",
        "\n",
        "plt.figure(figsize=(10, 6))\n",
        "plt.plot(effects, beta, marker='o')\n",
        "plt.title('Зависимость ошибки второго рода от эффекта')\n",
        "plt.xlabel('Эффект')\n",
        "plt.ylabel('Ошибка второго рода (β)')\n",
        "plt.grid(True)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 564
        },
        "id": "xrU9cflwcwj3",
        "outputId": "4af42be8-85fd-4fd0-9484-1ca94cc78f18"
      },
      "id": "xrU9cflwcwj3",
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Видим, что чем большее значение эффекта, тем больше ошибка второго рода, но график сильно волатильный.\n",
        "По идее, чем больше эффект, тем меньше должна быть ошибка второго рода, потому что разница между группами будет больше с ростом эффекта и вероятность сказать, что они не отличаются (при том, что они по факту отличаются) должна стать меньше. То есть в идеале получить убывающий график, наверное."
      ],
      "metadata": {
        "id": "2DKj8jUThYuO"
      },
      "id": "2DKj8jUThYuO"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "755a793e-5b49-4b13-85c6-830bc8f876ee"
      },
      "source": [
        "### Задание 7 (1.5 балл)\n",
        "\n",
        "На метрике revenue per user для разных групп (файл `ab_test_logs.pkl`) посчитайте мощность t-теста, бутстрапа и критерия Манна-Уитни при разных ошибках первого рода. Постройте график зависимости мощности от О1Р.\n",
        "\n",
        "Какой тест получился мощнее остальных на уровне значимости 5%?"
      ],
      "id": "755a793e-5b49-4b13-85c6-830bc8f876ee"
    },
    {
      "cell_type": "code",
      "source": [
        "def calculate_power(rev_control, rev_test, alpha, n_iter=1000, bootstrap=False, ttest=False, mann=False, sample_size=1000):\n",
        "    power = 0\n",
        "    for _ in range(n_iter):\n",
        "        if bootstrap:\n",
        "            a_sample = np.random.choice(rev_control, size=sample_size, replace=True)\n",
        "            b_sample = np.random.choice(rev_test, size=sample_size, replace=True)\n",
        "            _, p_value = stats.ttest_ind(a_sample, b_sample)\n",
        "            if p_value < alpha:\n",
        "                power += 1\n",
        "        elif ttest:\n",
        "            _, p_value = stats.ttest_ind(rev_first, rev_second)\n",
        "            if p_value < alpha:\n",
        "                power += 1\n",
        "        else:\n",
        "            _, p_value = stats.mannwhitneyu(rev_first, rev_second)\n",
        "            if p_value < alpha:\n",
        "                power += 1\n",
        "    return power / n_iter\n",
        "\n",
        "alphas = np.linspace(0.01, 0.2, 10)\n",
        "\n",
        "powers_ttest = [calculate_power(rev_control, rev_test, alpha, ttest=True) for alpha in alphas]\n",
        "powers_bootstrap = [calculate_power(rev_control, rev_test, alpha, bootstrap=True, sample_size=1000, n_iter=1000) for alpha in alphas]\n",
        "powers_mannwhitney = [calculate_power(rev_control, rev_test, alpha, mann=True) for alpha in alphas]\n",
        "\n",
        "plt.figure(figsize=(10, 6))\n",
        "plt.plot(alphas, powers_ttest, label='T-test')\n",
        "plt.plot(alphas, powers_bootstrap, label='Bootstrap')\n",
        "plt.plot(alphas, powers_mannwhitney, label='Mann-Whitney')\n",
        "plt.legend()\n",
        "plt.xlabel('Ошибка первого рода (О1Р)')\n",
        "plt.ylabel('Мощность')\n",
        "plt.title('Зависимость мощности от О1Р')\n",
        "plt.grid()\n",
        "plt.show()\n",
        "\n",
        "alpha = 0.05\n",
        "ttest = calculate_power(rev_control, rev_test, alpha, ttest=True)\n",
        "bootstrap = calculate_power(rev_control, rev_test, alpha, bootstrap=True, sample_size=1000, n_iter=1000)\n",
        "mannwhitney = calculate_power(rev_control, rev_test, alpha, mann=True)\n",
        "\n",
        "print(f\"Мощность t-теста: {ttest}\")\n",
        "print(f\"Мощность бутстрапа: {bootstrap}\")\n",
        "print(f\"Мощность Манна-Уитни: {mannwhitney}\")\n",
        "\n",
        "most_powerful = max(ttest, bootstrap, mannwhitney)\n",
        "print(\"Самый мощный тест:\")\n",
        "if most_powerful == ttest:\n",
        "    print(\"T-test\")\n",
        "elif most_powerful == bootstrap:\n",
        "    print(\"Bootstrap\")\n",
        "else:\n",
        "    print(\"Mann-Whitney U\")\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 655
        },
        "id": "mfcPmM2ujC2U",
        "outputId": "299a920f-d61d-4de5-fc48-0ad379a7d768"
      },
      "id": "mfcPmM2ujC2U",
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Мощность t-теста: 0.0\n",
            "Мощность бутстрапа: 0.27\n",
            "Мощность Манна-Уитни: 0.0\n",
            "Самый мощный тест:\n",
            "Bootstrap\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Наиболее мощный тест в нашем случае получился бутстрап. Возможно это связано со свойствами бутстрапа, при большом числе итераций мы получаем несмещенную оценку. Ну или все намного проще - я просто облажался :("
      ],
      "metadata": {
        "id": "QCc-U25YYl03"
      },
      "id": "QCc-U25YYl03"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "d95ebac0-19e1-49b6-929e-06a6bea46dc8"
      },
      "source": [
        "### Задание 8 (2 балла)"
      ],
      "id": "d95ebac0-19e1-49b6-929e-06a6bea46dc8"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "b1f6e544-f4e8-455c-828a-9ef8aa572cbe"
      },
      "source": [
        "На данных `aa_test_logs.pkl` реализуйте А/А тесты через бутстрап и постройте распределение p-value. Какое распределение у вас получилось? Почему? Попробуйте доказать этот факт"
      ],
      "id": "b1f6e544-f4e8-455c-828a-9ef8aa572cbe"
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "def bootstrap_test(a, b, n_bootstrap=10000, sample_size=None, effect=0):\n",
        "    \"\"\"Проводим бутстрап-тестирование, возвращаем список p-value.\"\"\"\n",
        "    pvalues = []\n",
        "    for _ in range(n_bootstrap):\n",
        "        a_sample = np.random.choice(a, size=sample_size, replace=True)\n",
        "        b_with_effect = b + effect\n",
        "        b_sample = np.random.choice(b_with_effect, size=sample_size, replace=True)\n",
        "        t_stat, pvalue = stats.ttest_ind(a_sample, b_sample)\n",
        "        pvalues.append(pvalue)\n",
        "\n",
        "    return np.array(pvalues)\n",
        "\n",
        "# Пример использования функции\n",
        "pvalues_bootstrap = bootstrap_test(rev_first, rev_second, n_bootstrap=1000, sample_size=1000, effect=0)\n",
        "\n",
        "# Визуализация результатов\n",
        "plt.hist(pvalues_bootstrap, bins=50, color='red', edgecolor='black')\n",
        "plt.xlabel('P-value')\n",
        "plt.ylabel('Частота')\n",
        "plt.title('Гистограмма P-value бутстрап-тестирования')\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "id": "bzkntgVyXGdI",
        "outputId": "6c6cf55e-ecfe-4125-87c8-287901925198"
      },
      "id": "bzkntgVyXGdI",
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "def plot_pvalue_distribution(pvalues_aa, alpha, beta):\n",
        "    \"\"\"Рисует графики распределения p-value.\"\"\"\n",
        "    estimated_first_type_error = np.mean(pvalues_aa < alpha)\n",
        "    y_one = estimated_first_type_error\n",
        "    X = np.linspace(0, 1, 1000)\n",
        "    Y_aa = [np.mean(pvalues_aa < x) for x in X]\n",
        "\n",
        "    plt.plot(X, Y_aa, label='A/A')\n",
        "    plt.plot([alpha, alpha], [0, 1], '--k', alpha=0.8)\n",
        "    plt.plot([0, alpha], [y_one, y_one], '--k', alpha=0.8)\n",
        "    plt.plot([0, 1], [0, 1], '--k', alpha=0.8)\n",
        "\n",
        "    plt.title('Оценка распределения p-value', size=16)\n",
        "    plt.xlabel('p-value', size=12)\n",
        "    plt.legend(fontsize=12)\n",
        "    plt.grid()\n",
        "    plt.show()\n",
        "\n",
        "plot_pvalue_distribution(pvalues_bootstrap, alpha, beta)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 478
        },
        "id": "dcjY8IMbdJ9-",
        "outputId": "acc3022d-0e95-4d19-81ef-c4f7b9b325a4"
      },
      "id": "dcjY8IMbdJ9-",
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "65da5521"
      },
      "source": [
        "Распределение p-value получилось равномерным, потому что мы использовали бутстрап, который при большом числе итераций выдает несмещенную оценку статистики. Как раз графики подтверждают это: распределение получилось равномерным и график проходит вдоль биссектрисы I координатного угла, а точка, соответствующая О1Р также лежит на этой прямой, значит критерий корректный"
      ],
      "id": "65da5521"
    }
  ],
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.9.12"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 5
}