Linear Regression Model (Python, SKlearn)
Objective
The objective of this project is to develop a predictive linear regression model to analyze store sales performance and identify key revenue drivers. The model's accuracy will be evaluated using performance metrics including MAPE (Mean Absolute Percentage Error) along with other relevant statistical measures. The analysis will deliver actionable insights to support data-driven business decisions.
Report
Through linear regression analysis, we identified 'Total Store Sq Ft', 'Centre Type Outlet', and 'Climate Hot' as statistically significant drivers of sales, supported by their coefficients and low p-values. The final model excluded 'Centre Type Strip' due to its negative impact. Variable selection involved:
Addressed multicollinearity by eliminating redundant variables
Iteratively refined the model using p-value significance
Validated robustness through MAPE and residual analysis
Confirmed 'Centre Type_Strip' had both negative impact and statistical significance
The final model balances predictive power with business interpretability, providing actionable insights for expansion strategy.