01 - Project - Perceptron.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "1b8fe662",
   "metadata": {},
   "source": [
    "# Project - Perceptron\n",
    "- Try a Perceptron model with more dimensions"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1c6c8002",
   "metadata": {},
   "source": [
    "### Step 1: Import libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2e1c1753",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.linear_model import Perceptron\n",
    "from sklearn import metrics\n",
    "from sklearn.model_selection import train_test_split\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "23b3991d",
   "metadata": {},
   "source": [
    "### Step 2: Read the data\n",
    "- Use Pandas [read_csv](https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.read_csv.html) method to read **files/weather.csv**\n",
    "- HINT: Use **parse_dates=True** and **index_col=0**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d3d98048",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "83ff3247",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "276d6763",
   "metadata": {},
   "source": [
    "### Step 3: Investigate data\n",
    "- Look for missing data points\n",
    "- You can do that by applying **isna()** and **sum()**, which will give a summary of rows missing data for each column.\n",
    "    - Resource: [isna()](https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.isna.html)\n",
    "    - Resource: [sum()](https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.sum.html)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ff620c74",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "32f0c679",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "01a32b51",
   "metadata": {},
   "source": [
    "### Step 4: Remove 'dirty' columns \n",
    "- Make a choice and remove columns with too many entries with NaN.\n",
    "- Say, take all columns with more than 100 rows.\n",
    "- Also, you can remove rows with non-numeric values (remember to keep **RainTomorrow**)\n",
    "- To remove rows use [drop(columns, axis=1)](https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.drop.html)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "738c0684",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b35c77d9",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "ac473dd7",
   "metadata": {},
   "source": [
    "### Step 5: Deal with remaining missing data\n",
    "- A simple choice is to simply remove rows with missing data\n",
    "- Use [dropna()](https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.dropna.html)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5cda89a8",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "65e7d67c",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "5db49d8e",
   "metadata": {},
   "source": [
    "### Step 6: Create training and test datasets\n",
    "- Define dataset **X** to consist of all data except **'RainTomorrow'**.\n",
    "    - Use [dropna()](https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.dropna.html)\n",
    "- Define dataset **y** to be datset cosisting of **'RainTomorrow'**.\n",
    "- Divide into **X_train, X_test, y_train, y_test** with **train_test_split**\n",
    "    - You can use **random_state=42** (or any other number) if you want to reproduce results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bfaa2a7a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a60b572b",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "809734e0",
   "metadata": {},
   "source": [
    "### Step 7: Train and test the model\n",
    "- Create classifier with **Perceptron**\n",
    "    - You can use **random_state=0** to be able to reproduce\n",
    "- Fit the model with training data **(X_train, y_train**)\n",
    "- Predict data from **X_test** (use predict) and assign to **y_pred**.\n",
    "- Evalute score by using **metrics.accuracy_score(y_test, y_pred)**.\n",
    "- You can redo with different choice of columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0d4243cc",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ffdae258",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "56f51ce7",
   "metadata": {},
   "source": [
    "### Step 8 (Optional): Plot the result\n",
    "- Use Matplotlib.pyplot (**plt**) with **subplots** to create a figure and axes (**fig, ax**)\n",
    "- Predict all the datapoints in **X**.\n",
    "- Make a scatter plot with all datapoints in **X** with color by the predictions made.\n",
    "    - You might want to use **alpha=0.25** in your plot as argument."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e7da5e81",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6d5f74e9",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4a55dfcc",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "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": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}