EazyML Modeling: Walmart Regression

Define Imports

!pip install --upgrade eazyml-automl
!pip install gdown python-dotenv

import os
import pandas as pd
import eazyml as ez
import gdown
from dotenv import load_dotenv
load_dotenv()

True

1. Initialize EazyML

The ez_init function uses the EAZYML_ACCESS_KEY environment variable for authentication. If the variable is not set, it defaults to a trial license.

ez.ez_init(os.getenv('EAZYML_ACCESS_KEY'))

{'success': True,
 'message': 'Initialized successfully. You may revoke your consent to sharing usage stats anytime. You have exclusive paid access.'}

2. Define Dataset Files, Outcome Variable and Train Model

2.1 Define Train Dataset and Other Model Parameters

gdown.download_folder(id='16LfwRMjchrPgdbsgPHr79AHvNCHsL5Is')

reg_file_path = os.path.join('data', "walmart_train_data.csv")
reg_outcome = "Weekly_Sales"
df_reg = pd.read_csv(reg_file_path)
options = {'model_type': 'predictive'}

2.2 Train Model

resp_reg = ez.ez_build_model(df_reg, outcome=reg_outcome, options=options)

2.3 Show Model Performance

ez.ez_display_df(resp_reg['model_performance'])

	Model	Rsquared	RMSE
0	Gradient Boosting Regressor	0.94	1835.38
1	Bagged Decision Trees	0.93	2000.09
2	Random Forest	0.93	2006.82
3	Boosted Decision Trees	0.76	3646.40
4	Lasso Regression	0.10	6950.23
5	Linear Regression	0.10	6950.26
6	Ridge Regression	0.10	6950.26

3. Dataset Information

The dataset used in this notebook is the Walmart Dataset, which contains data related to sales at Walmart stores. It includes various features such as store, fuel price, sales data, and other metrics over a specified period of time.

You can find more details and download the dataset from Kaggle using the following link:

Kaggle Walmart Dataset

Columns in the Dataset:

• Store: The store number.
• Weekly_Sales: Sales for the given store.
• IsHoliday: Whether the week is a special holiday week 1 – Holiday week 0 – Non-holiday week.
• Temperature: Temperature on the day of sale.
• Fuel_Price: Cost of fuel in the region.
• CPI: Prevailing consumer price index.
• Unemployment: Prevailing unemployment rate,

3.1 Display the Dataset

Below is a preview of the dataset:

# Load the dataset from the provided file
train = pd.read_csv(reg_file_path)

# Display the first few rows of the dataset
ez.ez_display_df(train.head())

	Store	Weekly_Sales	IsHoliday	Temperature	Fuel_Price	CPI	Unemployment
0	1	22516.313699	0.000000	42.310000	2.572000	211.096358	8.106000
1	1	22804.964444	1.000000	38.510000	2.548000	211.242170	8.106000
2	1	22081.755753	0.000000	39.930000	2.514000	211.289143	8.106000
3	1	19579.549861	0.000000	46.630000	2.561000	211.319643	8.106000
4	1	21298.721644	0.000000	46.500000	2.625000	211.350143	8.106000

4. Define Test Dataset and Predict on that Dataset

reg_model_info = resp_reg["model_info"]

4.1 Define Test Dataset

reg_test_file_path = os.path.join('data', "walmart_train_data.csv")
reg_test_data = pd.read_csv(reg_test_file_path)

4.2 Predict on Test Dataset

options = {}
reg_pred_df = ez.ez_predict(reg_test_data, model_info=reg_model_info, options=options)
pred_df = reg_pred_df['pred_df']

ez.ez_display_df(pred_df.head())

	Store	Weekly_Sales	IsHoliday	Temperature	Fuel_Price	CPI	Unemployment	Predicted Weekly_Sales
0	1	22516.313699	0	42.310000	2.572000	211.096358	8.106000	22386.510000
1	1	22804.964444	1	38.510000	2.548000	211.242170	8.106000	22894.660000
2	1	22081.755753	0	39.930000	2.514000	211.289143	8.106000	22262.500000
3	1	19579.549861	0	46.630000	2.561000	211.319643	8.106000	20111.170000
4	1	21298.721644	0	46.500000	2.625000	211.350143	8.106000	21138.440000