Rossmann store: Sales prediction

Abstract

The project features as a part of an active competition on Kaggle.com, which serves as a platform for various competitions on Analytics and Predictive Modelling. The purpose of the project is to forecast the sales for Rossmann, a drug store chain operating more than 3000 stores across 7 European countries and is the second largest in Germany. The main objective is to predict the daily sales for 1115 stores of Rossmann located across Germany. The sales forecast is for the next 7 weeks and the store sales are influenced by factors like promotions, seasonality, competition, locality, school and state holidays, etc. Based on the historical sales data for 31 months, a robust prediction model is to be built that accurately predicts the future sales and increases the productivity. To build a prediction model using store, promotion and competitor data to predict the sales 7 weeks in advance. The complete data set that has been provided for analysis is divided into 3 subsets – Training Dataset, Test Dataset and Store Dataset.