{ "cells": [ { "cell_type": "markdown", "id": "a221b1b1", "metadata": {}, "source": [ "# High Dimensional Fixed Effects with Rossman Sales Data\n", "\n", "**Intro**\n", "\n", "This tutorial demonstrates how to create models with high dimensional fixed effects using `glum`. Using `tabmat`, we can pass categorical variables with a large range of values. `glum` and `tabmat` will handle the creation of the one-hot-encoded design matrix.\n", "\n", "In some real-world problems, we have used millions of categories. This would be impossible with a dense matrix. General-purpose sparse matrices like compressed sparse row (CSR) matrices help but still leave a lot on the table. For a categorical matrix, we know that each row has only a single non-zero value and that value is 1. These optimizations are implemented in `tabmat.CategoricalMatrix`.\n", "\n", "\n", "**Background**\n", "\n", "For this tutorial, we will be predicting sales for the European drug store chain Rossman. Specifically, we are tasked with predicting daily sales for future dates. Ideally, we want a model that can capture the many factors that influence stores sales -- promotions, competition, school, holidays, seasonality, etc. As a baseline, we will start with a simple model that only uses a few basic predictors. Then, we will fit a model with a large number of fixed effects. For both models, we will use OLS with L2 regularization.\n", "\n", "We will use a gamma distribution for our model. This choice is motivated by two main factors. First, our target variable, sales, is a positive real number, which matches the support of the gamma distribution. Second, it is expected that factors influencing sales are multiplicative rather than additive, which is better captured with a gamma regression than say, OLS.\n", "\n", "\n", "*Note*: a few parts of this tutorial utilize local helper functions outside this notebook. If you wish to run the notebook on your own, you can find the rest of the code [here](https://github.com/Quantco/glum/tree/open-sourcing/docs/tutorials/rossman).\n", "\n", "## Table of Contents\n", "* [1. Data Loading and Feature Engineering](#1.-Data-Loading-and-Feature-Engineering)\n", "* [2. Fit Baseline GLM](#2.-Fit-baseline-GLM)\n", "* [3. GLM with High Dimensional Fixed Effects](#3.-GLM-with-High-Dimensional-Fixed-Effects)\n", "* [4. Plot Results](#4.-Plot-Results)" ] }, { "cell_type": "code", "execution_count": 1, "id": "2cd6396d", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "DataTransformerRegistry.enable('json')" ] }, "execution_count": 1, "metadata": {}, "output_type": "execute_result" } ], "source": [ "import os\n", "from pathlib import Path\n", "\n", "import altair as alt\n", "import matplotlib.pyplot as plt\n", "import numpy as np\n", "import pandas as pd\n", "\n", "from dask_ml.impute import SimpleImputer\n", "from dask_ml.preprocessing import Categorizer\n", "from glum import GeneralizedLinearRegressor\n", "from sklearn.pipeline import Pipeline\n", "\n", "from feature_engineering import apply_all_transformations\n", "from process_data import load_test, load_train, process_data\n", "\n", "import sys\n", "sys.path.append(\"../\")\n", "from metrics import root_mean_squared_percentage_error\n", "\n", "pd.set_option(\"display.float_format\", lambda x: \"%.3f\" % x)\n", "pd.set_option('display.max_columns', None)\n", "alt.data_transformers.enable(\"json\") # to allow for large plots" ] }, { "cell_type": "markdown", "id": "fc617222", "metadata": {}, "source": [ "## 1. Data loading and feature engineering\n", "[back to table of contents](#Table-of-Contents)\n", "\n", "We start by loading in the raw data. If you have not yet processed the raw data, it will be done below. (Initial processing consists of some basic cleaning and renaming of columns.\n", "\n", "*Note*: if you wish to run this notebook on your own, and have not done so already, please download the data from the [Rossman Kaggle Challenge](https://www.kaggle.com/c/rossmann-store-sales). This tutorial expects that it in a folder named \"raw_data\" under the same directory as the notebook." ] }, { "cell_type": "markdown", "id": "be095647", "metadata": {}, "source": [ "### 1.1 Load" ] }, { "cell_type": "code", "execution_count": 2, "id": "6a612662", "metadata": {}, "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", "
storeday_of_weekdatesalescustomersopenpromostate_holidayschool_holidayyearmonthstore_typeassortmentcompetition_distancecompetition_open_since_monthcompetition_open_since_yearpromo2promo2_since_weekpromo2_since_yearpromo_interval
1016095122013-01-0100False0a120131ca1270.0009.0002008.0000NaNNaNNone
1014980132013-01-025530668True00120131ca1270.0009.0002008.0000NaNNaNNone
1013865142013-01-034327578True00120131ca1270.0009.0002008.0000NaNNaNNone
1012750152013-01-044486619True00120131ca1270.0009.0002008.0000NaNNaNNone
1011635162013-01-054997635True00120131ca1270.0009.0002008.0000NaNNaNNone
\n", "
" ], "text/plain": [ " store day_of_week date sales customers open promo \\\n", "1016095 1 2 2013-01-01 0 0 False 0 \n", "1014980 1 3 2013-01-02 5530 668 True 0 \n", "1013865 1 4 2013-01-03 4327 578 True 0 \n", "1012750 1 5 2013-01-04 4486 619 True 0 \n", "1011635 1 6 2013-01-05 4997 635 True 0 \n", "\n", " state_holiday school_holiday year month store_type assortment \\\n", "1016095 a 1 2013 1 c a \n", "1014980 0 1 2013 1 c a \n", "1013865 0 1 2013 1 c a \n", "1012750 0 1 2013 1 c a \n", "1011635 0 1 2013 1 c a \n", "\n", " competition_distance competition_open_since_month \\\n", "1016095 1270.000 9.000 \n", "1014980 1270.000 9.000 \n", "1013865 1270.000 9.000 \n", "1012750 1270.000 9.000 \n", "1011635 1270.000 9.000 \n", "\n", " competition_open_since_year promo2 promo2_since_week \\\n", "1016095 2008.000 0 NaN \n", "1014980 2008.000 0 NaN \n", "1013865 2008.000 0 NaN \n", "1012750 2008.000 0 NaN \n", "1011635 2008.000 0 NaN \n", "\n", " promo2_since_year promo_interval \n", "1016095 NaN None \n", "1014980 NaN None \n", "1013865 NaN None \n", "1012750 NaN None \n", "1011635 NaN None " ] }, "execution_count": 2, "metadata": {}, "output_type": "execute_result" } ], "source": [ "if not all(Path(p).exists() for p in [\"raw_data/train.csv\", \"raw_data/test.csv\", \"raw_data/store.csv\"]):\n", " raise Exception(\"Please download raw data into 'raw_data' folder\")\n", "\n", "if not all(Path(p).exists() for p in [\"processed_data/train.parquet\", \"processed_data/test.parquet\"]):\n", " \"Processed data not found. Processing data from raw data...\"\n", " process_data()\n", " \"Done\"\n", "\n", "df = load_train().sort_values([\"store\", \"date\"])\n", "df = df.iloc[:int(.1*len(df))]\n", "df.head()" ] }, { "cell_type": "markdown", "id": "de96db23", "metadata": {}, "source": [ "### 1.2 Feature engineering\n", "\n", "As mentioned earlier, we want our model to incorporate many factors that could influence store sales. We create a number of fixed effects to capture this information. These include fixed effects for:\n", "\n", "- A certain number days before a school or state holiday\n", "- A certain number days after a school or state holiday\n", "- A certain number days before a promo\n", "- A certain number days after a promo\n", "- A certain number days before the store is open or closed\n", "- A certain number days after the store is open or closed\n", "- Each month for each store\n", "- Each year for each store\n", "- Each day of the week for each store\n", "\n", "We also do several other transformations like computing the z score to eliminate outliers (in the next step)" ] }, { "cell_type": "code", "execution_count": 3, "id": "59ff7624", "metadata": {}, "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", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
storeday_of_weekdatesalescustomersopenpromostate_holidayschool_holidayyearmonthstore_typeassortmentcompetition_distancecompetition_open_since_monthcompetition_open_since_yearpromo2promo2_since_weekpromo2_since_yearpromo_intervalage_quantilecompetition_opencountopen_lag_1open_lag_2open_lag_3open_lead_1open_lead_2open_lead_3promo_lag_1promo_lag_2promo_lag_3promo_lead_1promo_lead_2promo_lead_3school_holiday_lag_1school_holiday_lag_2school_holiday_lag_3school_holiday_lead_1school_holiday_lead_2school_holiday_lead_3state_holiday_lag_1state_holiday_lag_2state_holiday_lag_3state_holiday_lead_1state_holiday_lead_2state_holiday_lead_3store_day_of_weekstore_monthstore_school_holidaystore_state_holidaystore_yearzscore
1016095122013-01-0100False0a120131ca1270.0009.0002008.0000NaNNaNNone-11.00001.01.01.0TrueTrueTrue0.0000.0000.0000.0000.0000.0000.0000.0000.0001.0001.0001.0000000001_21_11_11_True1_2013NaN
1014980132013-01-025530668True00120131ca1270.0009.0002008.0000NaNNaNNone-11.0001False1.01.0TrueTrueTrue0.0000.0000.0000.0000.0000.0001.0000.0000.0001.0001.0001.000a000001_31_11_11_False1_2013NaN
1013865142013-01-034327578True00120131ca1270.0009.0002008.0000NaNNaNNone-11.0002TrueFalse1.0TrueTrueFalse0.0000.0000.0000.0000.0000.0001.0001.0000.0001.0001.0001.0000a00001_41_11_11_False1_2013NaN
1012750152013-01-044486619True00120131ca1270.0009.0002008.0000NaNNaNNone-11.0003TrueTrueFalseTrueFalseTrue0.0000.0000.0000.0000.0001.0001.0001.0001.0001.0001.0001.00000a0001_51_11_11_False1_2013NaN
1011635162013-01-054997635True00120131ca1270.0009.0002008.0000NaNNaNNone-11.0004TrueTrueTrueFalseTrueTrue0.0000.0000.0000.0001.0001.0001.0001.0001.0001.0001.0001.0000000001_61_11_11_False1_2013NaN
\n", "
" ], "text/plain": [ " store day_of_week date sales customers open promo \\\n", "1016095 1 2 2013-01-01 0 0 False 0 \n", "1014980 1 3 2013-01-02 5530 668 True 0 \n", "1013865 1 4 2013-01-03 4327 578 True 0 \n", "1012750 1 5 2013-01-04 4486 619 True 0 \n", "1011635 1 6 2013-01-05 4997 635 True 0 \n", "\n", " state_holiday school_holiday year month store_type assortment \\\n", "1016095 a 1 2013 1 c a \n", "1014980 0 1 2013 1 c a \n", "1013865 0 1 2013 1 c a \n", "1012750 0 1 2013 1 c a \n", "1011635 0 1 2013 1 c a \n", "\n", " competition_distance competition_open_since_month \\\n", "1016095 1270.000 9.000 \n", "1014980 1270.000 9.000 \n", "1013865 1270.000 9.000 \n", "1012750 1270.000 9.000 \n", "1011635 1270.000 9.000 \n", "\n", " competition_open_since_year promo2 promo2_since_week \\\n", "1016095 2008.000 0 NaN \n", "1014980 2008.000 0 NaN \n", "1013865 2008.000 0 NaN \n", "1012750 2008.000 0 NaN \n", "1011635 2008.000 0 NaN \n", "\n", " promo2_since_year promo_interval age_quantile competition_open \\\n", "1016095 NaN None -1 1.000 \n", "1014980 NaN None -1 1.000 \n", "1013865 NaN None -1 1.000 \n", "1012750 NaN None -1 1.000 \n", "1011635 NaN None -1 1.000 \n", "\n", " count open_lag_1 open_lag_2 open_lag_3 open_lead_1 open_lead_2 \\\n", "1016095 0 1.0 1.0 1.0 True True \n", "1014980 1 False 1.0 1.0 True True \n", "1013865 2 True False 1.0 True True \n", "1012750 3 True True False True False \n", "1011635 4 True True True False True \n", "\n", " open_lead_3 promo_lag_1 promo_lag_2 promo_lag_3 promo_lead_1 \\\n", "1016095 True 0.000 0.000 0.000 0.000 \n", "1014980 True 0.000 0.000 0.000 0.000 \n", "1013865 False 0.000 0.000 0.000 0.000 \n", "1012750 True 0.000 0.000 0.000 0.000 \n", "1011635 True 0.000 0.000 0.000 0.000 \n", "\n", " promo_lead_2 promo_lead_3 school_holiday_lag_1 \\\n", "1016095 0.000 0.000 0.000 \n", "1014980 0.000 0.000 1.000 \n", "1013865 0.000 0.000 1.000 \n", "1012750 0.000 1.000 1.000 \n", "1011635 1.000 1.000 1.000 \n", "\n", " school_holiday_lag_2 school_holiday_lag_3 school_holiday_lead_1 \\\n", "1016095 0.000 0.000 1.000 \n", "1014980 0.000 0.000 1.000 \n", "1013865 1.000 0.000 1.000 \n", "1012750 1.000 1.000 1.000 \n", "1011635 1.000 1.000 1.000 \n", "\n", " school_holiday_lead_2 school_holiday_lead_3 state_holiday_lag_1 \\\n", "1016095 1.000 1.000 0 \n", "1014980 1.000 1.000 a \n", "1013865 1.000 1.000 0 \n", "1012750 1.000 1.000 0 \n", "1011635 1.000 1.000 0 \n", "\n", " state_holiday_lag_2 state_holiday_lag_3 state_holiday_lead_1 \\\n", "1016095 0 0 0 \n", "1014980 0 0 0 \n", "1013865 a 0 0 \n", "1012750 0 a 0 \n", "1011635 0 0 0 \n", "\n", " state_holiday_lead_2 state_holiday_lead_3 store_day_of_week \\\n", "1016095 0 0 1_2 \n", "1014980 0 0 1_3 \n", "1013865 0 0 1_4 \n", "1012750 0 0 1_5 \n", "1011635 0 0 1_6 \n", "\n", " store_month store_school_holiday store_state_holiday store_year \\\n", "1016095 1_1 1_1 1_True 1_2013 \n", "1014980 1_1 1_1 1_False 1_2013 \n", "1013865 1_1 1_1 1_False 1_2013 \n", "1012750 1_1 1_1 1_False 1_2013 \n", "1011635 1_1 1_1 1_False 1_2013 \n", "\n", " zscore \n", "1016095 NaN \n", "1014980 NaN \n", "1013865 NaN \n", "1012750 NaN \n", "1011635 NaN " ] }, "execution_count": 3, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df = apply_all_transformations(df)\n", "df.head()" ] }, { "cell_type": "markdown", "id": "8209d0cd", "metadata": {}, "source": [ "### 1.3 Train vs. validation selection\n", "\n", "Lastly, we split our data into training and validation sets. Kaggle provides a test set for the Rossman challenge, but it does not directly include outcome data (sales), so we do not use it for our tutorial. Instead, we simulate predicting future sales by taking the last 5 months of our training data as our validation set." ] }, { "cell_type": "code", "execution_count": 4, "id": "f26974fd", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "(57876, 12502)" ] }, "execution_count": 4, "metadata": {}, "output_type": "execute_result" } ], "source": [ "validation_window = [pd.to_datetime(\"2015-03-15\"), pd.to_datetime(\"2015-07-31\")]\n", "select_train = (df[\"sales\"].gt(0) & df[\"date\"].lt(validation_window[0]) & df[\"zscore\"].abs().lt(5)).to_numpy()\n", "\n", "select_val = (\n", " df[\"sales\"].gt(0)\n", " & df[\"date\"].ge(validation_window[0])\n", " & df[\"date\"].lt(validation_window[1])\n", ").to_numpy()\n", "\n", "\n", "(select_train.sum(), select_val.sum())" ] }, { "cell_type": "markdown", "id": "bfa94a75", "metadata": {}, "source": [ "## 2. Fit baseline GLM\n", "[back to table of contents](#Table-of-Contents)\n", "\n", "We start with a simple model that uses only year, month, day of the week, and store as predictors. Even with these variables alone, we should still be able to capture a lot of valuable information. Year can capture overall sales trends, month can capture seasonality, week day can capture the variation in sales across the week, and store can capture locality. We will treat these all as categorical variables.\n", "\n", "With the `GeneralizedLinearRegressor()` class, we can pass in `pandas.Categorical` variables directly without having to encode them ourselves. This is convenient, especially when we start adding more fixed effects. But it is very important that the categories are aligned between calls to `fit` and `predict`. One way of achieving this alignment is with a `dask_ml.preprocessing.Categorizer`. Note, however, that the `Categorizer` class fails to enforce category alignment if the input column is already a categorical data type.\n", "\n", "You can reference the [pandas documentation on Categoricals](https://pandas.pydata.org/pandas-docs/stable/user_guide/categorical.html) to learn more about how these data types work." ] }, { "cell_type": "code", "execution_count": 5, "id": "428ec410", "metadata": {}, "outputs": [], "source": [ "baseline_features = [\"year\", \"month\", \"day_of_week\", \"store\"]\n", "baseline_categorizer = Categorizer(columns=baseline_features)\n", "baseline_glm = GeneralizedLinearRegressor(\n", " family=\"gamma\",\n", " scale_predictors=True,\n", " l1_ratio=0.0,\n", " alphas=1e-1,\n", ")" ] }, { "cell_type": "markdown", "id": "85157278", "metadata": {}, "source": [ "Fit the model making sure to process the data frame with the Categorizer first and inspect the coefficients." ] }, { "cell_type": "code", "execution_count": 6, "id": "c0535461", "metadata": {}, "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", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
interceptyear__2013year__2014year__2015month__1month__2month__3month__4month__5month__6month__7month__8month__9month__10month__11month__12day_of_week__1day_of_week__2day_of_week__3day_of_week__4day_of_week__5day_of_week__6day_of_week__7store__1store__2store__3store__4store__5store__6store__7store__8store__9store__10store__11store__12store__13store__14store__15store__16store__17store__18store__19store__20store__21store__22store__23store__24store__25store__26store__27store__28store__29store__30store__31store__32store__33store__34store__35store__36store__37store__38store__39store__40store__41store__42store__43store__44store__45store__46store__47store__48store__49store__50store__51store__52store__53store__54store__55store__56store__57store__58store__59store__60store__61store__62store__63store__64store__65store__66store__67store__68store__69store__70store__71store__72store__73store__74store__75store__76store__77store__78store__79store__80store__81store__82store__83store__84store__85store__86store__87store__88store__89store__90store__91store__92store__93store__94store__95store__96store__97store__98store__99store__100store__101store__102store__103store__104store__105store__106store__107store__108store__109store__110store__111store__112
coefficient8.781-0.0100.0080.003-0.015-0.021-0.0190.0200.008-0.006-0.002-0.023-0.034-0.0300.0220.1160.0980.010-0.014-0.0140.017-0.1000.285-0.156-0.1360.0240.202-0.161-0.0870.158-0.0940.002-0.0810.1100.074-0.126-0.0880.0090.082-0.0200.001-0.0080.085-0.088-0.178-0.0880.1830.2740.0080.192-0.0980.059-0.101-0.051-0.2250.1250.1020.1980.1850.046-0.048-0.153-0.128-0.0850.2340.017-0.086-0.101-0.0940.039-0.2430.058-0.2080.0120.079-0.1090.099-0.1800.0470.259-0.017-0.0910.102-0.149-0.0090.0340.251-0.122-0.0350.0860.0810.1900.0020.171-0.189-0.173-0.023-0.0380.1720.077-0.266-0.0910.0770.0380.163-0.2480.3690.019-0.155-0.060-0.048-0.0470.130-0.059-0.047-0.0520.0380.084-0.1210.007-0.119-0.1110.1080.0690.038-0.1710.233-0.1750.1360.0580.284-0.006-0.1740.018-0.039
\n", "
" ], "text/plain": [ " intercept year__2013 year__2014 year__2015 month__1 \\\n", "coefficient 8.781 -0.010 0.008 0.003 -0.015 \n", "\n", " month__2 month__3 month__4 month__5 month__6 month__7 \\\n", "coefficient -0.021 -0.019 0.020 0.008 -0.006 -0.002 \n", "\n", " month__8 month__9 month__10 month__11 month__12 \\\n", "coefficient -0.023 -0.034 -0.030 0.022 0.116 \n", "\n", " day_of_week__1 day_of_week__2 day_of_week__3 day_of_week__4 \\\n", "coefficient 0.098 0.010 -0.014 -0.014 \n", "\n", " day_of_week__5 day_of_week__6 day_of_week__7 store__1 \\\n", "coefficient 0.017 -0.100 0.285 -0.156 \n", "\n", " store__2 store__3 store__4 store__5 store__6 store__7 \\\n", "coefficient -0.136 0.024 0.202 -0.161 -0.087 0.158 \n", "\n", " store__8 store__9 store__10 store__11 store__12 store__13 \\\n", "coefficient -0.094 0.002 -0.081 0.110 0.074 -0.126 \n", "\n", " store__14 store__15 store__16 store__17 store__18 store__19 \\\n", "coefficient -0.088 0.009 0.082 -0.020 0.001 -0.008 \n", "\n", " store__20 store__21 store__22 store__23 store__24 store__25 \\\n", "coefficient 0.085 -0.088 -0.178 -0.088 0.183 0.274 \n", "\n", " store__26 store__27 store__28 store__29 store__30 store__31 \\\n", "coefficient 0.008 0.192 -0.098 0.059 -0.101 -0.051 \n", "\n", " store__32 store__33 store__34 store__35 store__36 store__37 \\\n", "coefficient -0.225 0.125 0.102 0.198 0.185 0.046 \n", "\n", " store__38 store__39 store__40 store__41 store__42 store__43 \\\n", "coefficient -0.048 -0.153 -0.128 -0.085 0.234 0.017 \n", "\n", " store__44 store__45 store__46 store__47 store__48 store__49 \\\n", "coefficient -0.086 -0.101 -0.094 0.039 -0.243 0.058 \n", "\n", " store__50 store__51 store__52 store__53 store__54 store__55 \\\n", "coefficient -0.208 0.012 0.079 -0.109 0.099 -0.180 \n", "\n", " store__56 store__57 store__58 store__59 store__60 store__61 \\\n", "coefficient 0.047 0.259 -0.017 -0.091 0.102 -0.149 \n", "\n", " store__62 store__63 store__64 store__65 store__66 store__67 \\\n", "coefficient -0.009 0.034 0.251 -0.122 -0.035 0.086 \n", "\n", " store__68 store__69 store__70 store__71 store__72 store__73 \\\n", "coefficient 0.081 0.190 0.002 0.171 -0.189 -0.173 \n", "\n", " store__74 store__75 store__76 store__77 store__78 store__79 \\\n", "coefficient -0.023 -0.038 0.172 0.077 -0.266 -0.091 \n", "\n", " store__80 store__81 store__82 store__83 store__84 store__85 \\\n", "coefficient 0.077 0.038 0.163 -0.248 0.369 0.019 \n", "\n", " store__86 store__87 store__88 store__89 store__90 store__91 \\\n", "coefficient -0.155 -0.060 -0.048 -0.047 0.130 -0.059 \n", "\n", " store__92 store__93 store__94 store__95 store__96 store__97 \\\n", "coefficient -0.047 -0.052 0.038 0.084 -0.121 0.007 \n", "\n", " store__98 store__99 store__100 store__101 store__102 \\\n", "coefficient -0.119 -0.111 0.108 0.069 0.038 \n", "\n", " store__103 store__104 store__105 store__106 store__107 \\\n", "coefficient -0.171 0.233 -0.175 0.136 0.058 \n", "\n", " store__108 store__109 store__110 store__111 store__112 \n", "coefficient 0.284 -0.006 -0.174 0.018 -0.039 " ] }, "execution_count": 6, "metadata": {}, "output_type": "execute_result" } ], "source": [ "baseline_glm.fit(\n", " baseline_categorizer.fit_transform(df[select_train][baseline_features]),\n", " df.loc[select_train, \"sales\"]\n", ")\n", "\n", "pd.DataFrame(\n", " {'coefficient': np.concatenate(([baseline_glm.intercept_], baseline_glm.coef_))},\n", " index=['intercept'] + baseline_glm.feature_names_\n", ").T" ] }, { "cell_type": "markdown", "id": "06a2d5a0", "metadata": {}, "source": [ "And let's predict for our test set with the caveat that we will predict 0 for days when the stores are closed!" ] }, { "cell_type": "code", "execution_count": 7, "id": "2c16ed1e", "metadata": {}, "outputs": [], "source": [ "df.loc[lambda x: x[\"open\"], \"predicted_sales_baseline\"] = baseline_glm.predict(\n", " baseline_categorizer.fit_transform(df.loc[lambda x: x[\"open\"]][baseline_features])\n", ")\n", "\n", "df[\"predicted_sales_baseline\"] = df[\"predicted_sales_baseline\"].fillna(0)\n", "df[\"predicted_sales_baseline\"] = df[\"predicted_sales_baseline\"]" ] }, { "cell_type": "markdown", "id": "555abace", "metadata": {}, "source": [ "We use root mean squared percentage error (RMSPE) as our performance metric. (Useful for thinking about error relative to total sales of each store)." ] }, { "cell_type": "code", "execution_count": 8, "id": "40c7d54f", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Training Error: 27.96%\n", "Validation Error: 29.57%\n" ] } ], "source": [ "train_err = root_mean_squared_percentage_error(\n", " df.loc[select_train, \"sales\"], df.loc[select_train, \"predicted_sales_baseline\"]\n", ")\n", "val_err = root_mean_squared_percentage_error(\n", " df.loc[select_val, \"sales\"], df.loc[select_val, \"predicted_sales_baseline\"]\n", ")\n", "print(f'Training Error: {round(train_err, 2)}%')\n", "print(f'Validation Error: {round(val_err, 2)}%')" ] }, { "cell_type": "markdown", "id": "1255336d", "metadata": {}, "source": [ "The results aren't bad for a start, but we can do better :)" ] }, { "cell_type": "markdown", "id": "83f614d4", "metadata": {}, "source": [ "## 3. GLM with high dimensional fixed effects\n", "[back to table of contents](#Table-of-Contents)\n", "\n", "Now, we repeat a similar process to above, but, this time, we take advantage of the full range of categoricals we created in our data transformation step. Since we will create a very large number of fixed effects, we may run into cases where our validation data has categorical values not seen in our training data. In these cases, Dask-ML's `Categorizer` will output null values when transforming the validation columns to the categoricals that were created on the training set. To fix this, we add Dask-ML's [SimpleImputer](https://ml.dask.org/modules/generated/dask_ml.impute.SimpleImputer.html) to our pipeline." ] }, { "cell_type": "code", "execution_count": 9, "id": "82a3fffd", "metadata": {}, "outputs": [], "source": [ "highdim_features = [\n", " \"age_quantile\",\n", " \"competition_open\",\n", " \"open_lag_1\",\n", " \"open_lag_2\",\n", " \"open_lag_3\",\n", " \"open_lead_1\",\n", " \"open_lead_2\",\n", " \"open_lead_3\",\n", " \"promo_lag_1\",\n", " \"promo_lag_2\",\n", " \"promo_lag_3\",\n", " \"promo_lead_1\",\n", " \"promo_lead_2\",\n", " \"promo_lead_3\",\n", " \"promo\",\n", " \"school_holiday_lag_1\",\n", " \"school_holiday_lag_2\",\n", " \"school_holiday_lag_3\",\n", " \"school_holiday_lead_1\",\n", " \"school_holiday_lead_2\",\n", " \"school_holiday_lead_3\",\n", " \"school_holiday\",\n", " \"state_holiday_lag_1\",\n", " \"state_holiday_lag_2\",\n", " \"state_holiday_lag_3\",\n", " \"state_holiday_lead_1\",\n", " \"state_holiday_lead_2\",\n", " \"state_holiday_lead_3\",\n", " \"state_holiday\",\n", " \"store_day_of_week\",\n", " \"store_month\",\n", " \"store_school_holiday\",\n", " \"store_state_holiday\",\n", " \"store_year\",\n", "]\n", "highdim_categorizer = Pipeline([\n", " (\"categorize\", Categorizer(columns=highdim_features)),\n", " (\"impute\", SimpleImputer(strategy=\"most_frequent\"))\n", "])\n", "highdim_glm = GeneralizedLinearRegressor(\n", " family=\"gamma\",\n", " scale_predictors=True,\n", " l1_ratio=0.0, # only ridge\n", " alpha=1e-1,\n", ")" ] }, { "cell_type": "markdown", "id": "8d3bb963", "metadata": {}, "source": [ "For reference, we output the total number of predictors after fitting the model. We can see that the number getting a bit larger, so we don't print out the coefficients this time." ] }, { "cell_type": "code", "execution_count": 10, "id": "f5152150", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Number of predictors: 2771\n" ] } ], "source": [ "highdim_glm.fit(\n", " highdim_categorizer.fit_transform(df[select_train][highdim_features]),\n", " df.loc[select_train, \"sales\"]\n", ")\n", "\n", "print(f\"Number of predictors: {len(highdim_glm.feature_names_)}\")" ] }, { "cell_type": "code", "execution_count": 11, "id": "5e616b3a", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Training Error: 12.54%\n", "Validation Error: 13.28%\n" ] } ], "source": [ "df.loc[lambda x: x[\"open\"], \"predicted_sales_highdim\"] = highdim_glm.predict(\n", " highdim_categorizer.transform(df.loc[lambda x: x[\"open\"]][highdim_features]),\n", ")\n", "\n", "df[\"predicted_sales_highdim\"] = df[\"predicted_sales_highdim\"].fillna(0)\n", "df[\"predicted_sales_highdim\"] = df[\"predicted_sales_highdim\"]\n", "\n", "\n", "train_err = root_mean_squared_percentage_error(\n", " df.loc[select_train, \"sales\"], df.loc[select_train, \"predicted_sales_highdim\"]\n", ")\n", "val_err = root_mean_squared_percentage_error(\n", " df.loc[select_val, \"sales\"], df.loc[select_val, \"predicted_sales_highdim\"]\n", ")\n", "print(f'Training Error: {round(train_err, 2)}%')\n", "print(f'Validation Error: {round(val_err, 2)}%')" ] }, { "cell_type": "markdown", "id": "0ef14c7b", "metadata": {}, "source": [ "From just the RMSPE, we can see a clear improvement from our baseline model." ] }, { "cell_type": "markdown", "id": "11010f80", "metadata": {}, "source": [ "## 4. Plot results\n", "[back to table of contents](#Table-of-Contents)\n", "\n", "Finally, to get a better look at our results, we make some plots." ] }, { "cell_type": "code", "execution_count": 12, "id": "56c28443", "metadata": {}, "outputs": [], "source": [ "sales_cols = [\"sales\", \"predicted_sales_highdim\", \"predicted_sales_baseline\"]" ] }, { "cell_type": "markdown", "id": "828bec37", "metadata": {}, "source": [ "First, we plot true sales and the sales predictions from each model aggregated over month:" ] }, { "cell_type": "code", "execution_count": 13, "id": "0389918c", "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "_, axs = plt.subplots(2, 1, figsize=(16, 16))\n", "\n", "for i, select in enumerate([select_train, select_val]):\n", " ax = axs[i]\n", " df_plot_date = df[select].groupby(\n", " [\"year\", \"month\"]\n", " ).agg(\"sum\")[sales_cols].reset_index()\n", "\n", " year_month_date = df_plot_date['month'].map(str)+ '-' + df_plot_date['year'].map(str)\n", " df_plot_date['year_month'] = pd.to_datetime(year_month_date, format='%m-%Y').dt.strftime('%m-%Y')\n", " df_plot_date.drop(columns= [\"year\", \"month\"], inplace=True)\n", "\n", " df_plot_date.plot(x=\"year_month\", ax=ax)\n", " ax.set_xticks(range(len(df_plot_date)))\n", " ax.set_xticklabels(df_plot_date.year_month, rotation=45)\n", " ax.set_xlabel(\"date\")\n", " ax.set_ylabel(\"Total sales\")\n", " ax.grid(True, linestyle='-.')\n", "\n", "axs[0].set_title(\"Training Results: Total Sales per Month\")\n", "axs[1].set_title(\"Validation Results: Total Sales per Month\")\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "fbecf40e", "metadata": {}, "source": [ "We can also look at aggregate sales for a subset of stores. We select the first 20 stores and plot in order of increasing sales." ] }, { "cell_type": "code", "execution_count": 14, "id": "a5e137bc", "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "_, axs = plt.subplots(2, 1, figsize=(14, 12))\n", "for i, select in enumerate([select_train, select_val]):\n", " ax = axs[i]\n", " df_plot_store = df[select].groupby(\n", " [\"store\"]\n", " ).agg(\"sum\")[sales_cols].reset_index()[:20].sort_values(by=\"sales\")\n", "\n", " df_plot_store.plot.bar(x=\"store\", ax=ax)\n", " ax.set_xlabel(\"Store\")\n", " ax.set_ylabel(\"Total sales\")\n", "\n", "axs[0].set_title(\"Training Results: Total Sales by Store\")\n", "axs[1].set_title(\"Validation Results: Total Sales by Store\")\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "4012a43b", "metadata": {}, "source": [ "We can see that the high dimensional model is much better at capturing the variation between months and individual stores!" ] } ], "metadata": { "jupytext": { "formats": "ipynb,md" }, "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": "3.8.10" } }, "nbformat": 4, "nbformat_minor": 5 }