dataset_analysis.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Amazon Fine Food dataset analysis",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyON7Xo8DN/a0UT14kNHfcbO",
      "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/stephan-strate/sentiment-fine-foods/blob/master/dataset_analysis.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jwIhVjI_6Bqz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# ensure images directory exists\n",
        "from pathlib import Path\n",
        "Path(\"images\").mkdir(parents=True, exist_ok=True)"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3g1mkypcX3dy",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "import urllib.request\n",
        "import zipfile\n",
        "\n",
        "TU_GREEN = \"#639a00\"\n",
        "\n",
        "url = \"https://storage.statc.de/datasets/amazon-fine-food/amazon-fine-food.zip\"\n",
        "filehandle, _ = urllib.request.urlretrieve(url)\n",
        "zip_file_object = zipfile.ZipFile(filehandle, \"r\")\n",
        "zip_file_object.extractall()"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fRvtQ4Bj8EEf",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "9321308d-476b-4b2d-ff37-56eefea5f88a"
      },
      "source": [
        "df = pd.read_csv(\"amazon-fine-food.csv\", delimiter=\",\", header=0)\n",
        "print(\"Dataset has {:,} rows\".format(df.shape[0]))"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Dataset has 568,454 rows\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PMaVeb87jTFq",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "outputId": "a77a71d2-2ac0-4638-add5-bacd36bcbfcf"
      },
      "source": [
        "df[[\"HelpfulnessNumerator\", \"HelpfulnessDenominator\", \"Score\", \"Text\"]].sample(5)"
      ],
      "execution_count": 4,
      "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>HelpfulnessNumerator</th>\n",
              "      <th>HelpfulnessDenominator</th>\n",
              "      <th>Score</th>\n",
              "      <th>Text</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>164770</th>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>5</td>\n",
              "      <td>Wow does my dog Pie LOVE these treats!! I was ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>37892</th>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>5</td>\n",
              "      <td>I really enjoy this tea and buying it through ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>338052</th>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>5</td>\n",
              "      <td>I've been buying cat treats for many years. Of...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>251529</th>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>I bought this toor dal from SimplyBeautiful. I...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>255322</th>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>5</td>\n",
              "      <td>This has been my go-to coffee for decades. Yes...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "        HelpfulnessNumerator  ...                                               Text\n",
              "164770                     0  ...  Wow does my dog Pie LOVE these treats!! I was ...\n",
              "37892                      0  ...  I really enjoy this tea and buying it through ...\n",
              "338052                     0  ...  I've been buying cat treats for many years. Of...\n",
              "251529                     3  ...  I bought this toor dal from SimplyBeautiful. I...\n",
              "255322                     1  ...  This has been my go-to coffee for decades. Yes...\n",
              "\n",
              "[5 rows x 4 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "67Jji6LRXuCK",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "b34eb4dd-383e-4c45-d239-6a400aef3809"
      },
      "source": [
        "import re\n",
        "\n",
        "df_cleaned = df.copy()\n",
        "\n",
        "def cleanhtml(series):\n",
        "  raw_text = series.Text\n",
        "  cleanr = re.compile(\"<.*?>\")\n",
        "  clean_text = re.sub(cleanr, \" \", raw_text)\n",
        "  return clean_text\n",
        "\n",
        "# clean html content of texts\n",
        "df_cleaned[\"TextClean\"] = df_cleaned.apply(cleanhtml, axis=1)\n",
        "df_cleaned[\"Html\"] = df_cleaned.apply(lambda x: x.Text != x.TextClean, axis=1)\n",
        "print(\"Cleaned {:,} rows from html ({} %)\".format(df_cleaned.Html.sum(), round(df_cleaned.Html.sum() / len(df.index) * 100)))"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Cleaned 147,324 rows from html (26.0 %)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_94m9YCZB7rO",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "bc8fa80d-169b-4222-faff-d166fc5ef43e"
      },
      "source": [
        "# removing duplicates\n",
        "df_cleaned.drop_duplicates(subset=[\"TextClean\", \"Score\"], inplace=True)\n",
        "rows_removed = len(df.index) - len(df_cleaned.index)\n",
        "print(\"Removed {:,} duplicates ({} %)\".format(rows_removed, round(rows_removed / len(df.index) * 100)))"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Removed 174,784 duplicates (31 %)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dz5Kxwt7-i87",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "d7a9bd42-b8e8-40ef-ac20-a858842229aa"
      },
      "source": [
        "print(\"Dataset size after cleaning: {:,} / {:,} rows ({} %)\".format(len(df_cleaned.index), len(df.index), round(len(df_cleaned.index) / len(df.index) * 100)))"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Dataset size after cleaning: 393,670 / 568,454 rows (69 %)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EVvI93YUPY2L",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 385
        },
        "outputId": "0e34989b-bdf2-4694-ff13-be19bc275fc7"
      },
      "source": [
        "import statistics\n",
        "\n",
        "label_values = set(df_cleaned.Score.values)\n",
        "\n",
        "fig, axs = plt.subplots(1, 3, constrained_layout=True, figsize=(10, 5))\n",
        "\n",
        "score_distribution = axs[0]\n",
        "label_value_counts = df_cleaned.Score.value_counts()\n",
        "score_distribution.bar(list(map(str, label_values)), list(map(lambda x: label_value_counts.get(x) / len(df_cleaned.index), label_values)), color=TU_GREEN)\n",
        "score_distribution.set_title(\"Reviews\")\n",
        "score_distribution.set_xlabel(\"Scores\")\n",
        "score_distribution.set_ylabel(\"Relative count\")\n",
        "\n",
        "def mean_helpfulness(label):\n",
        "  subset = df_cleaned[df_cleaned.Score.eq(label)]\n",
        "  return statistics.mean(filter(lambda x: x is not None, map(lambda index: index.HelpfulnessNumerator / index.HelpfulnessDenominator if index.HelpfulnessDenominator > 0 else None, subset.itertuples(index=True))))\n",
        "\n",
        "helpfulness_relative_score = axs[1]\n",
        "helpfulness_relative_score.bar(list(map(str, label_values)), list(map(mean_helpfulness, label_values)), color=TU_GREEN)\n",
        "helpfulness_relative_score.set_title(\"Helpfulness\")\n",
        "helpfulness_relative_score.set_xlabel(\"Scores\")\n",
        "helpfulness_relative_score.set_ylabel(\"Mean\")\n",
        "\n",
        "def count_helpful(label):\n",
        "  subset = df_cleaned[df_cleaned.Score.eq(label)]\n",
        "  return sum(map(lambda index: index.HelpfulnessDenominator, subset.itertuples(index=True))) / len(subset.values)\n",
        "\n",
        "helpfulness_count_score = axs[2]\n",
        "helpfulness_count_score.bar(list(map(str, label_values)), list(map(count_helpful, label_values)), color=TU_GREEN)\n",
        "helpfulness_count_score.set_title(\"Helpfulness\")\n",
        "helpfulness_count_score.set_xlabel(\"Scores\")\n",
        "helpfulness_count_score.set_ylabel(\"Mean count / review\")\n",
        "\n",
        "plt.savefig(\"images/dataset_general_analysis.pdf\")\n",
        "plt.show()"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x360 with 3 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "slpL1NTcV-S0",
        "colab_type": "text"
      },
      "source": [
        "For BERT it is important to determine a plausible input length of word embeddings. That is why we need to take a look at some statistics about review length (word count).\n",
        "\n",
        "We will overestimate the input token length in this case, because the input will probably contain some unknown words and therefore the tokenizer will split words into multiple parts.\n",
        "\n",
        "**Length:** 128"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z5sC47QwPfd5",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "07df91f7-b233-4d51-987d-5c054a3aa486"
      },
      "source": [
        "groups = df_cleaned.groupby(\"Score\")\n",
        "\n",
        "plt.bar(list(map(str, label_values)), list(map(lambda x: np.mean(list(map(lambda y: len(y.split(\" \")), x[1].Text.values))), groups)), color=TU_GREEN)\n",
        "plt.xlabel(\"Score\")\n",
        "plt.ylabel(\"Mean word count\")\n",
        "plt.title(\"Review length by score\")\n",
        "\n",
        "plt.savefig(\"images/dataset_review_length.pdf\")\n",
        "plt.show()"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "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": "code",
      "metadata": {
        "id": "KCOiO3q-MMl1",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 312
        },
        "outputId": "6acd59bf-4efe-49e5-85f3-13e876a11258"
      },
      "source": [
        "word_counts = np.array(list(map(lambda x: len(x.split(\" \")), df_cleaned.Text.values)))\n",
        "\n",
        "plt.hist(word_counts, range=(0, 256), bins=24, color=TU_GREEN)\n",
        "plt.axvline(word_counts.mean(), color=\"k\", linestyle=\"dashed\")\n",
        "\n",
        "plt.xlabel(\"Word count\")\n",
        "plt.ylabel(\"Count\")\n",
        "plt.title(\"Review length distribution\")\n",
        "\n",
        "plt.savefig(\"images/dataset_review_length_distribution.pdf\")\n",
        "plt.show()\n",
        "\n",
        "print(\"Mean review word count: {:,.2f} tokens\".format(word_counts.mean()))"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "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"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Mean review word count: 81.51 tokens\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}