Code.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Table of Content\n",
    "- [Load Datasets](#load_data)\n",
    "- [Prepare Data](#prepare_data)\n",
    "    - [Unify & Concat Dataframes](#concat_data)\n",
    "    - [Create Testdata & Trainingdata](#split_data)\n",
    "    - [Tokenize Strings](#tokenize_strings)\n",
    "- [Training](#train_data)\n",
    "- [Deploy & Test data](#deploy_data)\n",
    "    - [Export Model](#export_model)\n",
    "    - [Have some fun](#fun)\n",
    "\n",
    "This project is based on Tensorflows tutorial for basic text classification:\n",
    "https://www.tensorflow.org/tutorials/keras/text_classification"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a id=\"load_data\"></a>\n",
    "# Load Datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "trump_df = pd.read_csv(\"./data/RealDonaldTrumpTweets.csv\")\n",
    "trump_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "biden_df = pd.read_csv(\"./data/JoeBidenTweets.csv\")\n",
    "biden_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "musk_df = pd.read_csv(\"./data/ElonMuskTweets.csv\")\n",
    "musk_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "kardashian_df = pd.read_csv(\"./data/KimKardashianTweets.csv\")\n",
    "kardashian_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a id=\"prepare_data\"></a>\n",
    "# Prepare Data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a id=\"concat_data\"></a>\n",
    "## Unify & Concat Dataframes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "trump_unified_df = trump_df.rename(columns={\"content\": \"tweet\"})\n",
    "trump_unified_df[\"author\"] = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "biden_unified_df = biden_df\n",
    "biden_unified_df[\"author\"] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "musk_unified_df = musk_df.rename(columns={\"Tweet\": \"tweet\"})\n",
    "musk_unified_df[\"author\"] = 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "kardashian_unified_df = kardashian_df.reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "kardashian_unified_df.rename(columns={\"text\": \"tweet\"}, inplace=True)\n",
    "kardashian_unified_df[\"author\"] = 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "features = [\"tweet\", \"author\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.concat([\n",
    "    trump_unified_df[features],\n",
    "    biden_unified_df[features],\n",
    "    musk_unified_df[features],\n",
    "    kardashian_unified_df[features]\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a id=\"split_data\"></a>\n",
    "## Trainingdata & Testdata Split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train, df_test = train_test_split(\n",
    "    df, test_size=0.15, shuffle=True\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a id=\"tokenize_strings\"></a>\n",
    "## Tokenize Strings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "\n",
    "from tensorflow.keras.layers.experimental.preprocessing import TextVectorization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "max_features = 10000\n",
    "sequence_length = 200"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "vectorize_layer = TextVectorization(\n",
    "    max_tokens=max_features,\n",
    "    output_mode='int',\n",
    "    output_sequence_length=sequence_length\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "vectorize_layer.adapt(df_train[\"tweet\"].values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# This execution will need some time do be finished!\n",
    "# It took me 5 minutes for that cell\n",
    "tokenized_training_strings = np.array([\n",
    "    vectorize_layer(\n",
    "        tf.expand_dims(tweet, -1)\n",
    "    )[0] for tweet in df_train[\"tweet\"].values\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenized_testing_strings = np.array([\n",
    "    vectorize_layer(\n",
    "        tf.expand_dims(tweet, -1)\n",
    "    )[0] for tweet in df_test[\"tweet\"].values\n",
    "])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a id=\"train_data\"></a>\n",
    "# Train Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras import layers\n",
    "from tensorflow.keras import losses"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "embedding_dim = 16"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = tf.keras.Sequential([\n",
    "    layers.Embedding(max_features + 1, embedding_dim),\n",
    "    layers.Dropout(0.2),\n",
    "    layers.GlobalAveragePooling1D(),\n",
    "    layers.Dropout(0.2),\n",
    "    layers.Dense(4)\n",
    "])\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(\n",
    "    loss=losses.SparseCategoricalCrossentropy(from_logits=True),\n",
    "    optimizer='adam',\n",
    "    metrics=['accuracy']\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "history = model.fit(\n",
    "    x=tokenized_training_strings,\n",
    "    y=df_train[\"author\"],\n",
    "    validation_data=(tokenized_testing_strings, df_test[\"author\"]),\n",
    "    epochs=epochs\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a id=\"deploy_data\"></a>\n",
    "# Deploy & Test Data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a id=\"export_model\"></a>\n",
    "## Export Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "export_model = tf.keras.Sequential([\n",
    "    vectorize_layer,\n",
    "    model,\n",
    "    layers.Activation('sigmoid')\n",
    "])\n",
    "\n",
    "export_model.compile(\n",
    "    loss=losses.SparseCategoricalCrossentropy(from_logits=False),\n",
    "    optimizer=\"adam\",\n",
    "    metrics=['accuracy']\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a id=\"fun\"></a>\n",
    "## Have some fun with Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "biden_tweet = \"\"\"\n",
    "    We're on the right track. The American Rescue Plan laid a strong foundation for a new economy that brings everybody along, but it's just the first step. We must pass the American Jobs Plan to build an economy that works for everyone.\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "musk_tweet = \"\"\"\n",
    "    0 to 60mph in under 2 secs. Quickest production car ever made of any kind. Has to be felt to be believed.\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "trump_tweet = \"\"\"\n",
    "    96% Approval Rating in the Republican Party. Thank you!\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "kardashian_tweet = \"\"\"\n",
    "    I’m so embarrassing about wanting to post that pic!!!! Should I post it on National Peach Day!!!????\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Remember:\n",
    "- 0: Trump\n",
    "- 1: Biden\n",
    "- 2: Musk\n",
    "- 3: Kardashian"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "export_model.predict([\n",
    "    biden_tweet,\n",
    "    musk_tweet,\n",
    "    trump_tweet,\n",
    "    kardashian_tweet\n",
    "])"
   ]
  }
 ],
 "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}