This document summarizes a Kaggle competition to predict Walmart sales numbers using linear regression models. The data included past weekly sales numbers, markdown events, and associated economic factors like CPI, temperature, and unemployment. Different techniques were tested to handle missing data and improve the models, including linear regression with weights for holiday weeks. The best results on the leaderboard used more advanced algorithms like autoregression and random forests. While linear regression provided a good starting point, more sophisticated methods may have benefited from the limited sales prediction data.