Rossmann Sales Prediction
Matheus Guerreiro
Published
0
Development of a sales forecasting model for the Rossmann pharmaceutical retail chain using time series regression algorithms, aimed at cost reduction. The model reduced the average forecasting error from 36% to 4.65% (-31%) for the next six weeks, representing approximately €19 million in additional monthly revenue, using Kaggle data.
Tools used:
Python, Pandas, NumPy, Seaborn, Scikit-learn, SciPy, and Boruta.
Anaconda and Jupyter Notebooks.
Machine Learning with XGBoost Regressor.
Render Cloud.
Flask API.
Telegram Bot API.