In the 1st part of the Capstone project I have created a predictive model that predicts the weekly sales of 45 stores for next 3 months. In the first part I haven’t considered the department level data and took the sum of department sales to find the total sales of each store.
In this part of the project I am extending the scope to department Level and trying to predict the sales of all the 98 departments for each store. Since my system couldn’t handle this massive amount of data I am selecting the first 5 stores data to implement the same. This is because by selecting the first 5 stores it would be a perfect prototype of the entire data.
Objective
Part1. Create a predictive model that predicts the weekly sales of 45 Stores (In this stage I am not planning to predict the sales on Department Level)
*Part2. Extend the project to Department Level (Future Work)*
Method
Here I have concentrated on part2 of the project. I have scaled down and took data of 5 stores which works exactly as a prototype for big data models. I have done feature engineering and created new feature from the existing features. The new feature is a combination of Store number and Department number. This is because while we convert to dummy variable the combination will have more explaining power on the data trends.
In this extension I haven’t considered the economic features or any other features. This is an attempt to predict the 3 months sales by considering timeline as another feature with the newly created.
Conclusion
I have come up with the models that predict the weekly sales of 98 departments in 5 stores with around 90% of accuracy for 3 months(90 Days). This have further scope to expand and check how it worked for each store or for each department. That would be the future development plans.
Check this in GitHub
