generate_residual_plot.py

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import os

from data_loader import DataLoader
from feature_engineering import FeatureEngineer
from forecast_model import ForecastModel

def generate_residual_plot():
    file_path = "Walmart.csv"
    if not os.path.exists(file_path):
        file_path = "../Walmart.csv"
    if not os.path.exists(file_path):
        # Fallback to temp_data.csv if available
        file_path = "temp_data.csv"
        
    print(f"Loading data from {file_path}...")
    loader = DataLoader(file_path)
    raw_df = loader.load_data()
    
    print("Engineering features...")
    fe = FeatureEngineer()
    processed_df = fe.create_features(raw_df)
    
    print("Splitting data...")
    X_train, X_test, y_train, y_test, feature_names = fe.split_data(processed_df)
    
    print("Training model...")
    model = ForecastModel()
    model.train(X_train[feature_names], y_train)
    
    print("Generating predictions...")
    y_pred = model.predict(X_test[feature_names])
    
    print("Calculating residuals...")
    residuals = y_test - y_pred
    
    print("Plotting residual plot...")
    plt.figure(figsize=(10, 6))
    plt.scatter(y_pred, residuals, alpha=0.6, s=30, color='teal')
    plt.axhline(y=0, color='red', linestyle='--', linewidth=2, label='Zero Error')
    plt.xlabel("Predicted Demand")
    plt.ylabel("Residuals (Actual - Predicted)")
    plt.title("Residual Plot of Demand Forecasting Model")
    plt.legend()
    plt.grid(True)
    
    print("Saving plot...")
    plt.tight_layout()
    plt.savefig('residual_plot.png', bbox_inches='tight')
    plt.close()
    print("Success: Plot saved to residual_plot.png")

if __name__ == "__main__":
    generate_residual_plot()