Project Summary - This project aimed to develop a predictive model for forecasting sales at Rossman stores using linear regression and regularization techniques. The dataset provided detailed information about store characteristics, promotional activities, holidays, and historical sales data.
The data exploration phase revealed important insights such as higher sales on weekends and the impact of promotions and holidays on sales. The relationship between competition distance and sales was also explored, indicating a potential influence of convenience and customer preference for clustered shopping areas.
The predictive model was built using linear regression and regularization techniques, specifically Lasso and Ridge regression. These techniques addressed issues such as overfitting and multicollinearity. The models were evaluated using metrics such as mean squared error, root mean squared error, mean absolute error, and R-squared score.