Martin_Lukan_Standardize_your_data.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Standardize your data.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/mlukan/GDA3B2021/blob/main/Martin_Lukan_Standardize_your_data.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "v6t-vvMpLhAJ"
      },
      "source": [
        "# Execute the code below"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lqNi8-C9KhCq",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 406
        },
        "outputId": "c812b8fd-1c68-40b2-af4b-03cd38bbd61d"
      },
      "source": [
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "\n",
        "url = \"https://raw.githubusercontent.com/murpi/wilddata/master/quests/spotify.zip\"\n",
        "df_music = pd.read_csv(url)\n",
        "df_zoom = df_music.loc[df_music.genre.isin(['Country', 'Classical']), ['genre', 'duration_ms', 'speechiness']].reset_index(drop = True)\n",
        "df_zoom"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<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>genre</th>\n",
              "      <th>duration_ms</th>\n",
              "      <th>speechiness</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Country</td>\n",
              "      <td>200013</td>\n",
              "      <td>0.0444</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Country</td>\n",
              "      <td>208187</td>\n",
              "      <td>0.0569</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Country</td>\n",
              "      <td>123360</td>\n",
              "      <td>0.0960</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Country</td>\n",
              "      <td>238600</td>\n",
              "      <td>0.0368</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Country</td>\n",
              "      <td>243000</td>\n",
              "      <td>0.0330</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17915</th>\n",
              "      <td>Country</td>\n",
              "      <td>179147</td>\n",
              "      <td>0.0322</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17916</th>\n",
              "      <td>Country</td>\n",
              "      <td>230400</td>\n",
              "      <td>0.0832</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17917</th>\n",
              "      <td>Country</td>\n",
              "      <td>216093</td>\n",
              "      <td>0.0268</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17918</th>\n",
              "      <td>Country</td>\n",
              "      <td>179947</td>\n",
              "      <td>0.0909</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17919</th>\n",
              "      <td>Country</td>\n",
              "      <td>238173</td>\n",
              "      <td>0.0362</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>17920 rows × 3 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "         genre  duration_ms  speechiness\n",
              "0      Country       200013       0.0444\n",
              "1      Country       208187       0.0569\n",
              "2      Country       123360       0.0960\n",
              "3      Country       238600       0.0368\n",
              "4      Country       243000       0.0330\n",
              "...        ...          ...          ...\n",
              "17915  Country       179147       0.0322\n",
              "17916  Country       230400       0.0832\n",
              "17917  Country       216093       0.0268\n",
              "17918  Country       179947       0.0909\n",
              "17919  Country       238173       0.0362\n",
              "\n",
              "[17920 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GUc_ToyALjdJ"
      },
      "source": [
        "# Standardization and classification\n",
        "\n",
        "You now have a dataset with Country and Classical musics, and 2 numerical features : duration and speechiness.\n",
        "Our goal will be to predict genre from numerical features."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZIbWjlGyMIV1"
      },
      "source": [
        "## Draw a scatterplot from df_zoom with\n",
        "- 'duration_ms' on X axis \n",
        "- 'speechiness' on Y axis\n",
        "- 'genre' in hue"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_BJLKhdjMY8N",
        "outputId": "4d2e80ec-7ce7-404d-e648-b16ab04798dd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 297
        }
      },
      "source": [
        "sns.scatterplot(data=df_zoom,x='duration_ms',y='speechiness',hue='genre')\n"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f8ff28cb610>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bWh5LxO6MjkA"
      },
      "source": [
        "## Classification\n",
        "From df_zoom : \n",
        "- define X (`duration_ms` and `speechiness`) \n",
        "- define y (`genre`)\n",
        "- split your datas into train and test datasets, and `random_state = 2`\n",
        "- perform 3 classification algorithms (Logistic Regression, KNN and Decision Tree)\n",
        "- score your 3 models with accuracy score on the train dataset and on the test dataset\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-qAYEO6QN6ta",
        "outputId": "d61cf88b-f616-406e-fe26-2f0d6dca9185",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "# Your code here :\n",
        "df_zoom.head(2)\n",
        "X=df_zoom.copy()\n",
        "X=X.iloc[:,1:]\n",
        "y=df_zoom.copy()\n",
        "y=y.iloc[:,0]\n",
        "\n",
        "\n",
        "# Your code here\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, y,random_state=2)\n",
        "from sklearn.neighbors import KNeighborsClassifier\n",
        "modelKNN = KNeighborsClassifier()\n",
        "modelKNN.fit(X_train, y_train)\n",
        "\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "modelLR=LogisticRegression()\n",
        "modelLR.fit(X_train, y_train)\n",
        "\n",
        "from sklearn.tree import DecisionTreeClassifier\n",
        "modelDTC = DecisionTreeClassifier()\n",
        "modelDTC.fit(X_train, y_train)\n",
        "\n",
        "print(f\"KNN accuracy on the test data set: {modelKNN.score(X_test,y_test)}\")\n",
        "\n",
        "print(f\"LR accuracy on the test data set: {modelLR.score(X_test,y_test)}\")\n",
        "\n",
        "print(f\"DTC accuracy on the test data set: {modelDTC.score(X_test,y_test)}\")\n"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "KNN accuracy on the test data set: 0.7129464285714285\n",
            "LR accuracy on the test data set: 0.5098214285714285\n",
            "DTC accuracy on the test data set: 0.7370535714285714\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "b9zPpxjhOWrJ"
      },
      "source": [
        "\n",
        "You have to find these accuracy scores for test set :\n",
        "- Logistic regression : 0.50982\n",
        "- KNN : 0.71295\n",
        "- Decision tree : 0.73728\n",
        "\n",
        "It seems Decision tree is better, but did you look at the overfitting ?"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Fqe6OCZrOg5f"
      },
      "source": [
        "## Standardization\n",
        "\n",
        "- Fit your scaler model on X_train\n",
        "- Transform X_train and X_test  with your scaler model into  X_train_scaled and X_test_scaled\n",
        "- perform and score the same 3 classification algorithms, but with X_train_scaled and X_test_scaled"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fO_T_Ma7PJR7",
        "outputId": "dff8b1f7-b807-47b9-f71a-9dff77a77037",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "from sklearn.preprocessing import StandardScaler\n",
        "# Create and fit a scaler model\n",
        "scaler = StandardScaler().fit(X_train)\n",
        "\n",
        "# Your scaler model can now transform your data\n",
        "X_train_scaled = scaler.transform(X_train)\n",
        "X_test_scaled = scaler.transform(X_test)\n",
        "\n",
        "from sklearn.neighbors import KNeighborsClassifier\n",
        "modelKNN1 = KNeighborsClassifier()\n",
        "modelKNN1.fit(X_train_scaled, y_train)\n",
        "\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "modelLR1=LogisticRegression()\n",
        "modelLR1.fit(X_train_scaled, y_train)\n",
        "\n",
        "from sklearn.tree import DecisionTreeClassifier\n",
        "modelDTC1 = DecisionTreeClassifier()\n",
        "modelDTC1.fit(X_train_scaled, y_train)\n",
        "\n",
        "print(f\"KNN accuracy on the scaled data set: {modelKNN1.score(X_test_scaled,y_test)}\")\n",
        "\n",
        "print(f\"LR accuracy on the scaled data set: {modelLR1.score(X_test_scaled,y_test)}\")\n",
        "\n",
        "print(f\"DTC accuracy on the scaled data set: {modelDTC1.score(X_test_scaled,y_test)}\")"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "KNN accuracy on the scaled data set: 0.7743303571428571\n",
            "LR accuracy on the scaled data set: 0.6667410714285714\n",
            "DTC accuracy on the scaled data set: 0.7370535714285714\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "g0ehPzz2PK3r"
      },
      "source": [
        "# Conclusion\n",
        "- Decision tree is insensitive to Standardization.\n",
        "- Logistic regression and KNN have better result after Standardization.\n",
        "\n",
        "\n",
        "We can remember that standardization is always good in preprocessing, before machine learning classifications and regressions. At worst, it does not change anything. At best, it improves results."
      ]
    }
  ]
}