Deep Multivariate Forecasting for E-Commerce Demand Prediction

The commercial value of demand forecasting has significantly rose in recent years due to e-commerce platforms that require highly-adaptable supply chains to meet consumer needs. In this work, we explored implementing a deep multivariate model for demand forecasting for a food delivery company in Denmark. We concluded a 57% accuracy improvement in demand forecasting of food delivery requests from our implementation, deployed live in 100 locations.