You are given 3 years of store-item sales data, and asked to predict 3 months of sales for 10 different items at 5 different restaurants. Your data set contains both structured and unstructured elements.

Where the data came from This data was created using a multi-agent simulation.

What is the best way to deal with seasonality? Should items be modeled separately, or can they be combined? Should you choose deep learning or ARIMA? Can either beat XGBoost?

This competition allows you to explore different models and improve your skills in forecasting.

Location: Ft. Lauderdale, FL, USA.

• Data from Jan 1, 2019 to Dec 31, 2021
• Dates only specify seasonality
• No black swan events: no pandemics, wars, market conditions, etc.
• Holidays matter.
• Data set contains "trends" and multiple seasonalities.