Data Study Group Final Report: CatsAi

Data Study Groups are week-long events at The Alan Turing Institute bringing together some of the country’s top talent from data science, artificial intelligence, and wider fields, to analyse real-world data science challenges.

Communicating high-street bakery sales predictions using counterfactual explanations

This challenge aims to help CatsAi better serve their client (a large wholesaler) to estimate bakery orders to reduce waste and under delivery. The main tasks were to predict high-street sales based on meteorological factors and apply explainability techniques to effectively communicate their outputs to the client.

The main research questions we wanted to answer were:

1. Which features are predictive of sales (i.e. orders placed to the warehouse)?
2. Powerful predictive models are often difficult to interpret. Can we explain which features are important to a business owner?