In the competitive retail industry, the ability to predict future sales accurately is crucial for operational and strategic planning. Product sales forecasting aims to estimate the number of products a store will sell in the future, based on various influencing factors such as store type, location, regional characteristics, promotional activities, and temporal variations (such as holidays and seasons). This project focuses on developing a predictive model that uses historical sales data from different stores to forecast sales for upcoming periods.
MAPE
MAPE was used for evaluating the Global and Region level forecasts.
MAE
MAE was used for evaluating the Store level forecasts.
All the code for the project can be found in the notebooks folder. These are further divided into the following subfolders: EDA and Modelling.
The Streamlit App can be accessed using the following link: https://sales-forecasting-gn.streamlit.app/
All the code can be found in the deploy folder.
https://www.gautamnaik.com/blog/sales-forecasting/
https://github.com/gautamnaik1994/SalesForecasting_ML_CaseStudy
https://www.youtube.com/watch?v=_q-_W6A3754&ab_channel=GautamNaik