Walmart-Demand-Prediction

This project uses the Facebook -Prophet model to perform Demand Forecasting for the M5-forecasting competition. The dataset involves unit sales of 3049 products, classified into 3 product categories and sold across 10 stores in 3 states. The Facebook prophet model was used to forecast demand at a product-store level 28 days ahead. Exploratory data analysis and timeseries validation was performed to identify and normalize regressors for Demand Prediction. Yearly, weekly and monthly seasonality of demand was identified and incorporated into the model. Additional details are in the article published in towards data science below – https://towardsdatascience.com/demand-forecasting-using-fb-prophet-e3d1444b9dd8