video1.0<iframe src="https://www.loom.com/embed/eead9e5a626e44cea48b951ccfc49251" frameborder="0" width="1920" height="1440" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>14401920Loomhttps://www.loom.com14401920https://cdn.loom.com/sessions/thumbnails/eead9e5a626e44cea48b951ccfc49251-a29b5108cf961356.gif348.459Hi, I am Sudhir Dubey, and this is my retail sales forecasting portfolio project. I loaded and checked the dataset, with 188,340 rows and 10 columns in train, and 22,265 rows and 8 columns in test, with no missing values or duplicates. I created calendar and encoded features, then ran EDA and tested 5 hypotheses on discount, holiday, store type, location type, and the orders to sales relationship. GradientBoosting performed best with MAE 8577.30, RMSE 12818.88, and RSOF 0.5816. I turned the model into a simple Flask app, built a Tableau dashboard, and prepared notebook, repo, blog, and deployment.