{ "cells": [ { "cell_type": "markdown", "metadata": { "id": "OTEIqhB7dwPq" }, "source": [ "# Pandas, Matplotlib, Plotlyを用いたデータ分析例\n", "POSレジ購買データ（疑似データ）を例に代表的な使い方を例示します。なおこれらの細かい使い方は膨大なため、今回は一例として紹介するに留めます。より詳細は参考サイトを参照ください（=調べながら使えるようになりましょう。個別相談には応じます）。\n", "- **大前提**\n", " - 今回は全体の流れを確認しながら一部コード修正して実行する練習をしていきます。より詳細なプログラムの書き方や実行方法はその都度説明しますので、今回で全てを覚える必要はありません。\n", "- 疑似データ\n", " - [data.csv](../assets/data.csv) を使います。\n", " - step 1: Google Colabratoryで新規ノートブックを用意してください。\n", " - step 2: 授業資料フォルダを開き、data.csvを用意してください。\n", " - step 3: data.csvをノートブックにアップロードしてください。\n", "- 今回のデータ説明\n", " - purchase_date: 購入日（yyyy-mm-dd hh:mm:ss）\n", " - year: 購入年 (yyyy)\n", " - month: 購入月 (mm)\n", " - day_of_week: 曜日 (str)\n", " - time_of_day: 購入日（hh:mm:ss）\n", " - weather: 天候 (str)\n", " - age_group: 年代 (str)\n", " - gender: 性別 (str), 10代〜40代、50代以上\n", " - persona: 顧客属性 (str), 社会人、学生\n", " - purchase_tendency: 購入傾向 (dict)\n", " - product_name: 購入カテゴリ (list[str])\n", " - price: 購入金額 (float)\n", " - **疑似データ**です。同じ時間帯に天候が変わってたり、秒単位で同じ履歴があったり、金額に小数点があったりします。夏に雪がふることもあります。世界中に展開してるコンビニチェーン店全ログぐらいに捉えてください。\n", "\n", "- より深く学ぶための参考サイト\n", " - pandas: [ [公式チュートリアル](https://pandas.pydata.org/docs/user_guide/10min.html) | [100 pandas puzzles](https://notebook.community/prashantas/MyDataScience/Python/100-pandas-puzzles-with-solutions) | [cheat sheet](https://pandas.pydata.org/Pandas_Cheat_Sheet.pdf) ]\n", " - plotly: [ [公式チュートリアル](https://plotly.com/python/getting-started/) | [basic charts](https://plotly.com/python/basic-charts/) | [cheat sheet](https://images.plot.ly/plotly-documentation/images/python_cheat_sheet.pdf) ]\n", " - matplotlib: [ [公式チュートリアル](https://matplotlib.org/stable/tutorials/index.html) | [examples](https://matplotlib.org/stable/gallery/index.html) | [cheat sheet](https://matplotlib.org/cheatsheets/) ]\n", "- 具体的なデータを利用したチュートリアル\n", " - [マンガと学ぶデータビジュアライゼーション](https://kakeami.github.io/viz-madb/index.html)" ] }, { "cell_type": "markdown", "metadata": { "id": "SiRvcR9EdwPt" }, "source": [ "## データ読み込み〜例示\n", "pd.read_csv等で読み込んだデータはpd.DataFrame型（データフレーム型）と呼ばれる。ノートブック内でデータフレーム形式で保存した変数名を書くと、きれいな表敬式で出力してくれる。通常の.pyファイルでprint関数により出力することもできるが、ノートブックと比べると見づらい。\n", "- [pd.read_csv](https://pandas.pydata.org/docs/reference/api/pandas.read_csv.html): CSVファイル読み込み用の関数。\n", "- [pd.read_excel](https://pandas.pydata.org/docs/reference/api/pandas.read_excel.html): Excelファイルを読み込むこともできる（今回は未使用）" ] }, { "cell_type": "code", "execution_count": 1, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "ZI50vX6JeGuN", "outputId": "bf8315af-3c82-4773-9e15-f7ccedbff2bd" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Requirement already satisfied: japanize_matplotlib in /usr/local/lib/python3.10/dist-packages (1.1.3)\n", "Requirement already satisfied: matplotlib in /usr/local/lib/python3.10/dist-packages (from japanize_matplotlib) (3.7.1)\n", "Requirement already satisfied: contourpy>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib->japanize_matplotlib) (1.1.0)\n", "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.10/dist-packages (from matplotlib->japanize_matplotlib) (0.11.0)\n", "Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib->japanize_matplotlib) (4.42.1)\n", "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib->japanize_matplotlib) (1.4.4)\n", "Requirement already satisfied: numpy>=1.20 in /usr/local/lib/python3.10/dist-packages (from matplotlib->japanize_matplotlib) (1.23.5)\n", "Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib->japanize_matplotlib) (23.1)\n", "Requirement already satisfied: pillow>=6.2.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib->japanize_matplotlib) (9.4.0)\n", "Requirement already satisfied: pyparsing>=2.3.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib->japanize_matplotlib) (3.1.1)\n", "Requirement already satisfied: python-dateutil>=2.7 in /usr/local/lib/python3.10/dist-packages (from matplotlib->japanize_matplotlib) (2.8.2)\n", "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.10/dist-packages (from python-dateutil>=2.7->matplotlib->japanize_matplotlib) (1.16.0)\n" ] } ], "source": [ "# matplotlibで日本語を使うためのライブラリをインストール\n", "!pip install japanize_matplotlib" ] }, { "cell_type": "markdown", "metadata": { "id": "JypqX3iuypGS" }, "source": [ "### ライブラリの読み込み\n", "プログラミングにおける「ライブラリ」とは、特定の機能をまとめたものを指すことが多いです。\n", "\n", "日常生活で例えると、例えば英文を読んでいるときに分からない単語が出てきたら書籍やアプリにより英文辞書を用意し、調べるでしょう。これに対し数学を勉強しているときに英文辞書は不要なので片付けておくでしょう。このように必要なときに取り出し、不要なときには片付けておくことで卓上スペースを有効活用することができます。\n", "\n", "プログラムにおける「ライブラリの読み込み」は、同じように効率良く使うための工夫です。一般的にPCでの処理はストレージに保存されているものをメモリに読み込み、CPUやGPUにより処理されます。メモリはストレージに対して高速に動作しますがその分高価なので容量が小さいです。このためメモリには必要な情報だけを置いておきたいです。これを実現しているのが「ライブラリの読み込み」です。\n", "\n", "- 今回利用するライブラリ\n", " - pandasは表形式（正確にはデータフレーム形式）のデータを処理するためのライブラリです。``as pd``と書いているのは、プログラム中で「pd」と省略して書くために別称を設定しています。\n", " - plotly, matplotlibは、グラフ描画のためのライブラリです。\n", " - ``%matplotlib inline``は、ノートブック固有の記述です。ノートブック中にグラフを描画するために書いています。\n", " - japanize_matplotlibは、matplotlibで日本語フォントを利用するためのライブラリです。\n", "\n", "- 参考\n", " - [IT用語辞典 > ソフトウェア > プログラミング > 開発環境 > ライブラリとは](https://www.sophia-it.com/content/%E3%83%A9%E3%82%A4%E3%83%96%E3%83%A9%E3%83%AA)" ] }, { "cell_type": "code", "execution_count": 2, "metadata": { "id": "dXhUaMwxykHo" }, "outputs": [], "source": [ "# ライブラリの読み込み\n", "import pandas as pd\n", "import plotly.express as px\n", "\n", "%matplotlib inline\n", "import matplotlib.pyplot as plt\n", "import japanize_matplotlib" ] }, { "cell_type": "markdown", "metadata": { "id": "GvMjRdwi1Tkt" }, "source": [ "### データの読み込み\n", "- ``pd.read_csv()`` は、「pdライブラリの中にある関数read_csv()を実行せよ」という命令です。カッコ内でCSVファイルの名前を指定することで、該当ファイルを読み込み、データフレームとして利用することができます。\n", "- ``df = pd.read_csv()`` は、「pd.read_csv()の実行結果を変数dfに保存せよ」という命令です。実行結果を変数に保存しておかないと、参照する（利用する）ことができません。\n", "- ``df`` は、「変数dfの中身を出力しろ」という命令です。" ] }, { "cell_type": "code", "execution_count": 3, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 779 }, "id": "DY3YU2C_dwPt", "outputId": "3e3c3911-3117-4145-cc68-71d9b8eb077a" }, "outputs": [ { "data": { "text/html": [ "\n", "

\n", "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	purchase_date	year	month	day_of_week	time_of_day	weather	age_group	gender	persona	purchase_tendency	product_name	price
0	2022-01-01 10:47:25	2022	1	Saturday	10:47:25	晴れ	10代	男性	学生	{'おにぎり': 0.1, '野菜飲料': 0.05, 'エナジードリンク': 0.1500...	['お菓子']	158.39
1	2022-01-01 11:20:06	2022	1	Saturday	11:20:06	雨	10代	男性	社会人	{'おにぎり': 0.1, '野菜飲料': 0.05, 'エナジードリンク': 0.05, ...	['傘']	1422.74
2	2022-01-01 11:43:07	2022	1	Saturday	11:43:07	雨	30代	女性	社会人	{'おにぎり': 0.2, '野菜飲料': 0.2, 'エナジードリンク': 0.25, '...	['エナジードリンク']	483.69
3	2022-01-01 11:58:43	2022	1	Saturday	11:58:43	晴れ	20代	女性	学生	{'おにぎり': 0.2, '野菜飲料': 0.2, 'エナジードリンク': 0.35, '...	['お菓子']	173.22
4	2022-01-01 12:11:03	2022	1	Saturday	12:11:03	晴れ	40代	男性	学生	{'おにぎり': 0.2, '野菜飲料': 0.25, 'エナジードリンク': 0.1500...	['アイス']	248.47
...	...	...	...	...	...	...	...	...	...	...	...	...
9995	2022-12-31 21:34:22	2022	12	Saturday	21:34:22	晴れ	50代以上	男性	学生	{'おにぎり': 0.3, '野菜飲料': 0.25, 'エナジードリンク': 0.1500...	['清涼飲料水']	784.93
9996	2022-12-31 21:55:54	2022	12	Saturday	21:55:54	晴れ	10代	男性	社会人	{'おにぎり': 0.1, '野菜飲料': 0.05, 'エナジードリンク': 0.05, ...	['お菓子']	152.61
9997	2022-12-31 22:03:36	2022	12	Saturday	22:03:36	晴れ	50代以上	女性	学生	{'おにぎり': 0.3, '野菜飲料': 0.4, 'エナジードリンク': 0.15000...	['お菓子']	131.72
9998	2022-12-31 22:25:55	2022	12	Saturday	22:25:55	晴れ	30代	男性	学生	{'おにぎり': 0.2, '野菜飲料': 0.05, 'エナジードリンク': 0.35, ...	['おにぎり']	1227.61
9999	2022-12-31 22:29:17	2022	12	Saturday	22:29:17	曇り	30代	男性	学生	{'おにぎり': 0.2, '野菜飲料': 0.05, 'エナジードリンク': 0.35, ...	['ビッグ弁当']	552.63

\n", "

10000 rows × 12 columns

\n", "

\n", "

\n", "\n", "

\n", " \n", "\n", " \n", "\n", " \n", "

\n", "\n", "\n", "

\n", " \n", "\n", "\n", "\n", " \n", "

\n", "

\n", "

\n" ], "text/plain": [ " purchase_date year month day_of_week time_of_day weather \\\n", "0 2022-01-01 10:47:25 2022 1 Saturday 10:47:25 晴れ \n", "1 2022-01-01 11:20:06 2022 1 Saturday 11:20:06 雨 \n", "2 2022-01-01 11:43:07 2022 1 Saturday 11:43:07 雨 \n", "3 2022-01-01 11:58:43 2022 1 Saturday 11:58:43 晴れ \n", "4 2022-01-01 12:11:03 2022 1 Saturday 12:11:03 晴れ \n", "... ... ... ... ... ... ... \n", "9995 2022-12-31 21:34:22 2022 12 Saturday 21:34:22 晴れ \n", "9996 2022-12-31 21:55:54 2022 12 Saturday 21:55:54 晴れ \n", "9997 2022-12-31 22:03:36 2022 12 Saturday 22:03:36 晴れ \n", "9998 2022-12-31 22:25:55 2022 12 Saturday 22:25:55 晴れ \n", "9999 2022-12-31 22:29:17 2022 12 Saturday 22:29:17 曇り \n", "\n", " age_group gender persona \\\n", "0 10代男性学生 \n", "1 10代男性社会人 \n", "2 30代女性社会人 \n", "3 20代女性学生 \n", "4 40代男性学生 \n", "... ... ... ... \n", "9995 50代以上男性学生 \n", "9996 10代男性社会人 \n", "9997 50代以上女性学生 \n", "9998 30代男性学生 \n", "9999 30代男性学生 \n", "\n", " purchase_tendency product_name price \n", "0 {'おにぎり': 0.1, '野菜飲料': 0.05, 'エナジードリンク': 0.1500... ['お菓子'] 158.39 \n", "1 {'おにぎり': 0.1, '野菜飲料': 0.05, 'エナジードリンク': 0.05, ... ['傘'] 1422.74 \n", "2 {'おにぎり': 0.2, '野菜飲料': 0.2, 'エナジードリンク': 0.25, '... ['エナジードリンク'] 483.69 \n", "3 {'おにぎり': 0.2, '野菜飲料': 0.2, 'エナジードリンク': 0.35, '... ['お菓子'] 173.22 \n", "4 {'おにぎり': 0.2, '野菜飲料': 0.25, 'エナジードリンク': 0.1500... ['アイス'] 248.47 \n", "... ... ... ... \n", "9995 {'おにぎり': 0.3, '野菜飲料': 0.25, 'エナジードリンク': 0.1500... ['清涼飲料水'] 784.93 \n", "9996 {'おにぎり': 0.1, '野菜飲料': 0.05, 'エナジードリンク': 0.05, ... ['お菓子'] 152.61 \n", "9997 {'おにぎり': 0.3, '野菜飲料': 0.4, 'エナジードリンク': 0.15000... ['お菓子'] 131.72 \n", "9998 {'おにぎり': 0.2, '野菜飲料': 0.05, 'エナジードリンク': 0.35, ... ['おにぎり'] 1227.61 \n", "9999 {'おにぎり': 0.2, '野菜飲料': 0.05, 'エナジードリンク': 0.35, ... ['ビッグ弁当'] 552.63 \n", "\n", "[10000 rows x 12 columns]" ] }, "execution_count": 3, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# データ読み込み\n", "# assets/data.csvをアップロードしてから実行する。\n", "df = pd.read_csv(\"data.csv\")\n", "df" ] }, { "cell_type": "markdown", "metadata": { "id": "R_pFGBptdwPu" }, "source": [ "## データフレームの概要\n", "列ごとのNullじゃない要素数や、数値データかどうかを確認する際には pd.info() が便利です。``Null``とは「欠落や欠損」を意味します。世の中には一部が欠落しているデータが良くあります。\n", "- [pd.info](https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.info.html)" ] }, { "cell_type": "code", "execution_count": 4, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "Jzvvs_aBdwPv", "outputId": "1cf5f38b-089e-4ca2-b850-dacb0458723f" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n", "RangeIndex: 10000 entries, 0 to 9999\n", "Data columns (total 12 columns):\n", " # Column Non-Null Count Dtype \n", "--- ------ -------------- ----- \n", " 0 purchase_date 10000 non-null object \n", " 1 year 10000 non-null int64 \n", " 2 month 10000 non-null int64 \n", " 3 day_of_week 10000 non-null object \n", " 4 time_of_day 10000 non-null object \n", " 5 weather 10000 non-null object \n", " 6 age_group 10000 non-null object \n", " 7 gender 10000 non-null object \n", " 8 persona 10000 non-null object \n", " 9 purchase_tendency 10000 non-null object \n", " 10 product_name 10000 non-null object \n", " 11 price 10000 non-null float64\n", "dtypes: float64(1), int64(2), object(9)\n", "memory usage: 937.6+ KB\n" ] } ], "source": [ "# 各列の基本情報\n", "df.info()" ] }, { "cell_type": "markdown", "metadata": { "id": "4Y2HIO1EdwPv" }, "source": [ "## 数値データに関する基本統計量\n", "今回のデータでは year, month, price が数値データになっている。これらに対する基本統計量を見るには pd.descriobe() を使おう\n", "- [pd.describe](https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.describe.html)" ] }, { "cell_type": "code", "execution_count": 5, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 297 }, "id": "l0TnKhkNdwPv", "outputId": "859604d6-7d97-4736-f588-35a9d768861a" }, "outputs": [ { "data": { "text/html": [ "\n", "

\n", "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	year	month	price
count	10000.0	10000.000000	10000.000000
mean	2022.0	6.583700	889.433363
std	0.0	3.441334	688.305266
min	2022.0	1.000000	50.080000
25%	2022.0	4.000000	276.595000
50%	2022.0	7.000000	767.475000
75%	2022.0	10.000000	1263.915000
max	2022.0	12.000000	2998.880000

\n", "

\n", "

\n", "\n", "

\n", " \n", "\n", " \n", "\n", " \n", "

\n", "\n", "\n", "

\n", " \n", "\n", "\n", "\n", " \n", "

\n", "

\n", "

\n" ], "text/plain": [ " year month price\n", "count 10000.0 10000.000000 10000.000000\n", "mean 2022.0 6.583700 889.433363\n", "std 0.0 3.441334 688.305266\n", "min 2022.0 1.000000 50.080000\n", "25% 2022.0 4.000000 276.595000\n", "50% 2022.0 7.000000 767.475000\n", "75% 2022.0 10.000000 1263.915000\n", "max 2022.0 12.000000 2998.880000" ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# 数値データに関する基本統計量\n", "df.describe()" ] }, { "cell_type": "code", "execution_count": 6, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "a_SWG6sV4zAo", "outputId": "34990a95-2e34-453e-a7b0-25c1fff43c9b" }, "outputs": [ { "data": { "text/plain": [ "2998.88" ] }, "execution_count": 6, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# 特定の値、例えば price列の max を抜き出すには次のように書く。\n", "# 書式：df['列名'].describe()['行名']\n", "df['price'].describe()['max']" ] }, { "cell_type": "markdown", "metadata": { "id": "viQAHatDdwPv" }, "source": [ "## 数値データの出現分布 => ヒストグラム\n", "ヒストグラムを描画するには matplotlib か plotly を使うことが多い。どちらかというとplotlyの方がスタイリッシュに見えやすくインタラクティブな図も作成しやすいことから、ここではplotlyを使ってみよう。\n", "\n", "描画ライブラリは指定できる引数が極めて多く、最低限データフレームとその列名を与える必要がある。\n", "- [px.histogram](https://plotly.com/python/histograms/)\n", " - 代表的な引数\n", " - x: データフレームにおける列名。\n", " - nbins: ビン数。\n", " - marginal: \"box\"と書くと箱ひげ図も追加。" ] }, { "cell_type": "code", "execution_count": 7, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 542 }, "id": "u6UqRKfBdwPw", "outputId": "763fed06-38a2-4a49-9182-07e91cf0fbbe" }, "outputs": [ { "data": { "text/html": [ "\n", "\n", "\n", "

\n", "

\n", "\n", "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# 数値データに関するヒストグラム\n", "px.histogram(df, x=\"price\")" ] }, { "cell_type": "code", "execution_count": 8, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 542 }, "id": "s9O0H3LodwPw", "outputId": "01464f5f-b131-4690-bc5f-f64f6900f696" }, "outputs": [ { "data": { "text/html": [ "\n", "\n", "\n", "

\n", "

\n", "\n", "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# パラメータ指定によりカスタマイズできる\n", "px.histogram(df, x=\"price\", nbins=20, marginal=\"box\")" ] }, { "cell_type": "markdown", "metadata": { "id": "i5WAVSzedwPw" }, "source": [ "## カテゴリデータの出現回数\n", "product_nameには購入カテゴリ名が書かれている。各カテゴリが出現した回数をカウントするには df[\"カテゴリ名\"].value_counts() という形式で確認することができる。カテゴリ名を省略して実行することもできるが、この場合には参照しづらいこともあるので使い分けよう。\n", "- [pd.value_counts](https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.value_counts.html)" ] }, { "cell_type": "code", "execution_count": 9, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "kkyH6WQOdwPw", "outputId": "52fc8c49-1556-45ef-c8cf-47691887546f" }, "outputs": [ { "data": { "text/plain": [ "day_of_week\n", "Monday 1530\n", "Friday 1464\n", "Tuesday 1451\n", "Wednesday 1444\n", "Saturday 1418\n", "Sunday 1368\n", "Thursday 1325\n", "dtype: int64" ] }, "execution_count": 9, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# カテゴリデータに関する出現回数\n", "df.value_counts(\"day_of_week\")" ] }, { "cell_type": "markdown", "metadata": { "id": "QtRutPFCdwPx" }, "source": [ "## 演習3.1: カテゴリデータの出現回数\n", "他のカテゴリデータ（age_group, gender, product_name）についても出現回数を確認してみよう。" ] }, { "cell_type": "markdown", "metadata": { "id": "577B322wdwPx" }, "source": [ "## クロス分析\n", "今回のデータには age_group（顧客の年代カテゴリ）も含まれている。そこで「10代の購入したカテゴリ集計」「20代の購入したカテゴリ集計」というように年代カテゴリごとに集計を見てみたい、というような2変数以上の組み合わせを考慮することを「クロス分析」と呼ぶ。このような場合には pd.crosstab か pd.groupby を使おう。\n", "- [pd.crosstab](https://pandas.pydata.org/docs/reference/api/pandas.crosstab.html)\n", "- [pd.groupby](https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.groupby.html)\n" ] }, { "cell_type": "code", "execution_count": 10, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 252 }, "id": "4YB6C5_sdwPx", "outputId": "b2362c7b-c047-41ff-8937-5e282fb62734" }, "outputs": [ { "data": { "text/html": [ "\n", "

\n", "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

product_name	['おにぎり']	['お菓子']	['アイス']	['エナジードリンク']	['ビッグ弁当']	['傘']	['清涼飲料水']	['野菜飲料']
age_group
10代	78	161	186	76	301	101	255	103
20代	183	183	205	261	145	128	185	125
30代	220	189	220	303	187	128	202	148
40代	225	192	243	98	77	161	341	365
50代以上	712	448	513	242	184	354	762	810

\n", "

\n", "

\n", "\n", "

\n", " \n", "\n", " \n", "\n", " \n", "

\n", "\n", "\n", "

\n", " \n", "\n", "\n", "\n", " \n", "

\n", "

\n", "

\n" ], "text/plain": [ "product_name ['おにぎり'] ['お菓子'] ['アイス'] ['エナジードリンク'] ['ビッグ弁当'] ['傘'] \\\n", "age_group \n", "10代 78 161 186 76 301 101 \n", "20代 183 183 205 261 145 128 \n", "30代 220 189 220 303 187 128 \n", "40代 225 192 243 98 77 161 \n", "50代以上 712 448 513 242 184 354 \n", "\n", "product_name ['清涼飲料水'] ['野菜飲料'] \n", "age_group \n", "10代 255 103 \n", "20代 185 125 \n", "30代 202 148 \n", "40代 341 365 \n", "50代以上 762 810 " ] }, "execution_count": 10, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# クロス分析 (pd.groupbyの例)\n", "labels = [\"age_group\", \"product_name\"]\n", "data = df[labels]\n", "cross = data.groupby(labels).size().unstack() # 組み合わせの頻度を計算\n", "cross" ] }, { "cell_type": "code", "execution_count": 11, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 252 }, "id": "-pd5eMu130DN", "outputId": "affab439-1517-407a-af64-b5c18425eb15" }, "outputs": [ { "data": { "text/html": [ "\n", "

\n", "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

product_name	['おにぎり']	['お菓子']	['アイス']	['エナジードリンク']	['ビッグ弁当']	['傘']	['清涼飲料水']	['野菜飲料']
age_group
10代	78	161	186	76	301	101	255	103
20代	183	183	205	261	145	128	185	125
30代	220	189	220	303	187	128	202	148
40代	225	192	243	98	77	161	341	365
50代以上	712	448	513	242	184	354	762	810

\n", "

\n", "

\n", "\n", "

\n", " \n", "\n", " \n", "\n", " \n", "

\n", "\n", "\n", "

\n", " \n", "\n", "\n", "\n", " \n", "

\n", "

\n", "

\n" ], "text/plain": [ "product_name ['おにぎり'] ['お菓子'] ['アイス'] ['エナジードリンク'] ['ビッグ弁当'] ['傘'] \\\n", "age_group \n", "10代 78 161 186 76 301 101 \n", "20代 183 183 205 261 145 128 \n", "30代 220 189 220 303 187 128 \n", "40代 225 192 243 98 77 161 \n", "50代以上 712 448 513 242 184 354 \n", "\n", "product_name ['清涼飲料水'] ['野菜飲料'] \n", "age_group \n", "10代 255 103 \n", "20代 185 125 \n", "30代 202 148 \n", "40代 341 365 \n", "50代以上 762 810 " ] }, "execution_count": 11, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# クロス分析（pd.crosstabの例）\n", "cross = pd.crosstab(index=df[\"age_group\"], columns=df[\"product_name\"])\n", "cross\n" ] }, { "cell_type": "markdown", "metadata": { "id": "1hnvOcj7dwPx" }, "source": [ "## クロス分析結果を描画\n", "plotlyの棒グラフで描画するなら px.bar を使おう。\n", "- [px.bar](https://plotly.com/python/bar-charts/)" ] }, { "cell_type": "code", "execution_count": 12, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 542 }, "id": "X3cZWLIUdwPx", "outputId": "a40597e2-05c1-463f-d40f-c46c0c4df153" }, "outputs": [ { "data": { "text/html": [ "\n", "\n", "\n", "

\n", "

\n", "\n", "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# プロット作成（カウント）\n", "fig = px.bar(cross, barmode=\"stack\", text=\"product_name\")\n", "fig.show()" ] }, { "cell_type": "markdown", "metadata": { "id": "AhPIdRV4dwPx" }, "source": [ "## カウントではなく割合で見たい場合\n", "pd.sumで総和を取り、pd.divで割り算することで割合を求めた上で描画する。\n", "- [pd.sum](https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.sum.html)\n", "- [pd.div](https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.div.html)" ] }, { "cell_type": "code", "execution_count": 13, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 542 }, "id": "W-R2yOe6dwPy", "outputId": "d8b19cf8-dc13-4695-f6fe-c41a659e02b6" }, "outputs": [ { "data": { "text/html": [ "\n", "\n", "\n", "

\n", "

\n", "\n", "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# プロット作成（割合）\n", "# グループごとの合計値で各値を割る\n", "cross_ratio = cross.div(cross.sum(axis=1), axis=0)\n", "fig = px.bar(cross_ratio, barmode=\"stack\", text=\"product_name\")\n", "fig.show()" ] }, { "cell_type": "markdown", "metadata": { "id": "jJnwCd-FdwPy" }, "source": [ "## 演習3.2: クロス集計の描画\n", "以下のクロス集計を積み上げ棒グラフで作成し、気づいたことを報告しよう。\n", "- クロス1: gender x product_name\n", "- クロス2: persona x product_name\n", "- クロス3: month x product_name" ] }, { "cell_type": "markdown", "metadata": { "id": "fuZ-OAS4dwPy" }, "source": [ "## 3変数のクロス集計\n", "今回のデータには gender（性別）もある。「年代 x 性別」毎の集計を取るには、indexで複数指定する必要がある。" ] }, { "cell_type": "code", "execution_count": 14, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 407 }, "id": "ItpqCyBF5pGm", "outputId": "2e5f1c5a-5d36-47e2-b421-86fe8834ec54" }, "outputs": [ { "data": { "text/html": [ "\n", "

\n", "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	product_name	['おにぎり']	['お菓子']	['アイス']	['エナジードリンク']	['ビッグ弁当']	['傘']	['清涼飲料水']	['野菜飲料']
age_group	gender
10代	女性	33	69	80	36	125	47	127	83
男性	45	92	106	40	176	54	128	20
20代	女性	97	92	89	124	44	73	104	103
男性	86	91	116	137	101	55	81	22
30代	女性	104	90	116	139	55	65	87	122
男性	116	99	104	164	132	63	115	26
40代	女性	126	89	132	53	0	83	153	217
男性	99	103	111	45	77	78	188	148
50代以上	女性	341	241	255	120	0	183	399	505
男性	371	207	258	122	184	171	363	305

\n", "

\n", "

\n", "\n", "

\n", " \n", "\n", " \n", "\n", " \n", "

\n", "\n", "\n", "

\n", " \n", "\n", "\n", "\n", " \n", "

\n", "

\n", "

\n" ], "text/plain": [ "product_name ['おにぎり'] ['お菓子'] ['アイス'] ['エナジードリンク'] ['ビッグ弁当'] ['傘'] \\\n", "age_group gender \n", "10代女性 33 69 80 36 125 47 \n", " 男性 45 92 106 40 176 54 \n", "20代女性 97 92 89 124 44 73 \n", " 男性 86 91 116 137 101 55 \n", "30代女性 104 90 116 139 55 65 \n", " 男性 116 99 104 164 132 63 \n", "40代女性 126 89 132 53 0 83 \n", " 男性 99 103 111 45 77 78 \n", "50代以上女性 341 241 255 120 0 183 \n", " 男性 371 207 258 122 184 171 \n", "\n", "product_name ['清涼飲料水'] ['野菜飲料'] \n", "age_group gender \n", "10代女性 127 83 \n", " 男性 128 20 \n", "20代女性 104 103 \n", " 男性 81 22 \n", "30代女性 87 122 \n", " 男性 115 26 \n", "40代女性 153 217 \n", " 男性 188 148 \n", "50代以上女性 399 505 \n", " 男性 363 305 " ] }, "execution_count": 14, "metadata": {}, "output_type": "execute_result" } ], "source": [ "cross3 = pd.crosstab(index=[df[\"age_group\"],df[\"gender\"]], columns=df[\"product_name\"])\n", "cross3" ] }, { "cell_type": "markdown", "metadata": { "id": "r3hMhHQpdwPy" }, "source": [ "## 3変数クロス分析の描画\n", "残念ながら3変数クロス分析結果をplotlyでそのまま描画することはできない。工夫するとできるが、ここでは簡易さの観点から matplotlib を使ってみよう。\n", "- [plt.plot](https://matplotlib.org/stable/api/_as_gen/matplotlib.pyplot.plot.html)" ] }, { "cell_type": "code", "execution_count": 15, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 559 }, "id": "P6lfm_K8dwPy", "outputId": "99230df8-e546-44c2-c7bc-519c587e3c89" }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "

" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# クロス集計結果のカテゴリ順序を指定する\n", "age_order = sorted(list(df[\"age_group\"].unique()))\n", "cross3 = cross3.reindex(age_order, level='age_group')\n", "\n", "# バープロットを作成する\n", "ax = cross3.plot(kind='bar', stacked=True)\n", "\n", "# グラフの設定\n", "plt.xlabel('Age Group')\n", "plt.ylabel('Count')\n", "plt.title('Cross Tabulation')\n", "\n", "# 凡例を表示\n", "ax.legend(loc='center left', bbox_to_anchor=(1, 0.5))\n", "\n", "# プロットを表示する\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": { "id": "wCwgSbxL3Jcx" }, "source": [ "## 月毎の集計\n", "- [px.line](https://plotly.com/python/line-charts/)" ] }, { "cell_type": "code", "execution_count": 16, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 419 }, "id": "1VqzXmRr3Jcy", "outputId": "66791bb2-21f7-4809-aad9-2207f5920164" }, "outputs": [ { "data": { "text/html": [ "\n", "

\n", "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	month	product_name	purchase_count
0	1	['おにぎり']	104
1	1	['お菓子']	180
2	1	['アイス']	46
3	1	['エナジードリンク']	70
4	1	['ビッグ弁当']	84
...	...	...	...
91	12	['エナジードリンク']	72
92	12	['ビッグ弁当']	70
93	12	['傘']	67
94	12	['清涼飲料水']	146
95	12	['野菜飲料']	135

\n", "

96 rows × 3 columns

\n", "

\n", "

\n", "\n", "

\n", " \n", "\n", " \n", "\n", " \n", "

\n", "\n", "\n", "

\n", " \n", "\n", "\n", "\n", " \n", "

\n", "

\n", "

\n" ], "text/plain": [ " month product_name purchase_count\n", "0 1 ['おにぎり'] 104\n", "1 1 ['お菓子'] 180\n", "2 1 ['アイス'] 46\n", "3 1 ['エナジードリンク'] 70\n", "4 1 ['ビッグ弁当'] 84\n", ".. ... ... ...\n", "91 12 ['エナジードリンク'] 72\n", "92 12 ['ビッグ弁当'] 70\n", "93 12 ['傘'] 67\n", "94 12 ['清涼飲料水'] 146\n", "95 12 ['野菜飲料'] 135\n", "\n", "[96 rows x 3 columns]" ] }, "execution_count": 16, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# 商品別の購買数を月ごとに集計\n", "monthly_counts = df.groupby(['month', 'product_name']).size().reset_index(name='purchase_count')\n", "monthly_counts" ] }, { "cell_type": "code", "execution_count": 17, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 542 }, "id": "ne_3jNUS3Jcy", "outputId": "c0c6d234-bb9d-4403-e5c6-2b90cf23ee05" }, "outputs": [ { "data": { "text/html": [ "\n", "\n", "\n", "

\n", "

\n", "\n", "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# 購買数の集計結果をplotlyで描画\n", "fig = px.line(monthly_counts, x='month', y='purchase_count', color='product_name', title='Monthly Purchase Counts by Product')\n", "fig.show()" ] }, { "cell_type": "code", "execution_count": 17, "metadata": { "id": "XuRQx5yc3Jcy" }, "outputs": [], "source": [] } ], "metadata": { "colab": { "provenance": [] }, "kernelspec": { "display_name": "dm", "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": "3.9.6" }, "orig_nbformat": 4 }, "nbformat": 4, "nbformat_minor": 0 }