Ex2-3_plot_saled_prediction.ipynb

ex2-3_plot_saled_prediction.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Ex2-3_plot_saled_prediction.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyMeNqSBh9oOcCUuEddfbezz",
      "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/gist/chezou/d986ce3e908bb6e355ac4eaf97d23447/ex2-3_plot_saled_prediction.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "b5_NIT3Oag7A",
        "colab_type": "text"
      },
      "source": [
        "# 演習 2-3 売上データの予測結果をグラフに描画しよう\n",
        "\n",
        "Facebook Prophetで予測をした売上をJupyter notebookを使いプロットしてみましょう。"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SAC2PQKqbPin",
        "colab_type": "text"
      },
      "source": [
        "## pytdをインストールする\n",
        "\n",
        "このセルはローカル環境では実行しなくても構いませんが、Google Colaboratoryなどの環境では初回実行時に必ず実行してください。"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jUOS9-D-Z1UD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Install pytd\n",
        "!pip install -q pytd"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kswk78nEbbWG",
        "colab_type": "text"
      },
      "source": [
        "## TD API keyを設定する\n",
        "\n",
        "TDの読み書きができるmaster API keyを設定します。\n",
        "\n",
        "API keyは以下のURLから確認できます。\n",
        "https://console.treasuredata.com/app/mp/ak\n",
        "\n",
        "\"Input your Treasure Data API key\"の下のボックスにAPIキーをコピー＆ペーストをしてください。"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Pd5rPRW3ZnQX",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Set your TD API key\n",
        "import getpass\n",
        "\n",
        "print(\"Input your Treasure Data API key\")\n",
        "apikey = getpass.getpass()\n",
        "print(\"Succeeded\")"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vE4VfgoTcOQt",
        "colab_type": "text"
      },
      "source": [
        "## 各種設定を変更する\n",
        "\n",
        "下のセルで `db_name` の値を `timeseries` から `timeseries_YOURNAME` のようにYOURNAMEの部分を名前などに置き換えてください。\n",
        "また、USリージョン以外の場合は、 `endpoint` も適宜書き換えてください。\n",
        "\n",
        "- 東京: \"https://api.treasuredata.co.jp\"\n",
        "- EU: \"https://api.eu01.treasuredata.com\""
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b6YL8NJNZt8z",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Set configurations\n",
        "db_name = \"timeseries\"\n",
        "endpoint = \"https://api.treasuredata.com\""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MRr5djaHc5Td",
        "colab_type": "text"
      },
      "source": [
        "## グラフ描画用の関数を定義する\n",
        "\n",
        "次のセルではProphetのグラフを描画するための関数を定義しています。詳細は飛ばしてしまって構いません。"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "D3aRc87kZK-K",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "%matplotlib inline\n",
        "\n",
        "# Define function for plot\n",
        "\n",
        "from matplotlib import pyplot as plt\n",
        "from matplotlib.dates import (\n",
        "    AutoDateLocator,\n",
        "    AutoDateFormatter,\n",
        ")\n",
        "\n",
        "# Equivalent code with model.plot()\n",
        "# https://github.com/facebook/prophet/blob/ca9a49d328ab1f2a991f246a3ebc37a7f9c896c5/python/fbprophet/plot.py#L41-L88\n",
        "def plot(df, df_prev, xlabel='ds', ylabel='y'):\n",
        "    fig = plt.figure(facecolor=\"w\", figsize=(10,6))\n",
        "    ax = fig.add_subplot(111)\n",
        "\n",
        "    fcst_t = pd.to_datetime(df[\"ds\"])\n",
        "    ax.plot(pd.to_datetime(df_prev[\"ds\"]), df_prev[\"y\"], \"k.\")\n",
        "    ax.plot(fcst_t, df[\"yhat\"], ls=\"-\", c=\"#0072B2\")\n",
        "    if \"cap\" in df:\n",
        "        ax.plot(fcst_t, df[\"cap\"], ls=\"--\", c=\"k\")\n",
        "\n",
        "    if \"floor\" in df:\n",
        "        ax.plot(fcst_t, df[\"floor\"], ls=\"--\", c=\"k\")\n",
        "\n",
        "    ax.fill_between(fcst_t, df[\"yhat_lower\"], df[\"yhat_upper\"], color=\"#0072B2\", alpha=0.2)\n",
        "    locator = AutoDateLocator(interval_multiples=False)\n",
        "    formatter = AutoDateFormatter(locator)\n",
        "    ax.xaxis.set_major_locator(locator)\n",
        "    ax.xaxis.set_major_formatter(formatter)\n",
        "    ax.grid(True, which='major', c='gray', ls='-', lw=1, alpha=0.2)\n",
        "    ax.set_xlabel(xlabel)\n",
        "    ax.set_ylabel(ylabel)\n",
        "    fig.tight_layout()\n",
        "\n",
        "    return fig"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "buV3aaGkdCoY",
        "colab_type": "text"
      },
      "source": [
        "## TDから過去の売上データを取得する\n",
        "\n",
        "過去の売上データをpytdを使い取得します。また、取得しデータを折れ線グラフでプロットします。"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZrTi8NGYZSaH",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Load past sales data\n",
        "import pytd\n",
        "import pandas as pd\n",
        "\n",
        "pytd_client = pytd.Client(database=db_name, apikey=apikey, endpoint=endpoint)\n",
        "df_prev = pd.DataFrame(**pytd_client.query(\"select * from retail_sales order by ds\"))\n",
        "df_prev"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZwZR3dK_Za2a",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Plot past sales data\n",
        "df_prev.plot(x=\"ds\", y=\"y\", kind=\"line\", figsize=(10, 6))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1IBt5i0KdRP5",
        "colab_type": "text"
      },
      "source": [
        "## 予測結果をTDから取得する\n",
        "\n",
        "Custom Scripts環境で予測した結果をpytdで取得し、その結果をプロットします。"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ofL1wDVkZZr4",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Load predicted sales data\n",
        "df = pd.DataFrame(**pytd_client.query(\"select * from predicted_sales order by ds\"))\n",
        "df"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hjWipINnZd60",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Plot predicted sales data\n",
        "fig = plot(df, df_prev)"
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}

URL for binder: https://mybinder.org/v2/gist/chezou/d986ce3e908bb6e355ac4eaf97d23447/master?filepath=ex2-3_plot_saled_prediction.ipynb