Case Study: How do I increase sales and reduce waste?

Overview:

The client provides end-to-end retail solutions for airlines, one of which, is in-flight catering services. They were looking to make better use of an airline’s historic data to enable more accurate predictions of how much food will be required on flights, resulting in reduced wastage and a reduction of missed sales opportunities. The current catering provider to major Airlines was only achieving 50% accuracy with a manual process.

How did Heron approach this?

This project was to deliver a solution which could be used in making accurate catering predictions. Heron began this, by utilising the data that was available in order to make improved, more accurate predictions, straight away. We then processed the historical data from the airline in order to create an accurate prediction algorithm. The aim of this was to deliver a fully automated system which could extract the information from an email attachment from the airline, process it against an algorithm, (which had been trained using historical data), before sending out predictions to specified recipients.

The focus of this work was to deliver a solution which could be used in conjunction with the data and processes that they currently use, so as to start using the solution straight away. Following this we looked to evolve the solution by making use of additional data such as; sex, age groups, nationality and weather. This provided analytical reports as well as allowing for the fine tuning of the algorithm powering the solution.

Technologies Used

  • Machine Learning

The Outcome:

The outcome was a predictive analytics platform which provided ongoing insights and drove efficient planning of on-board stock levels to increase in sales and reduce waste. This facilitated ongoing improvements of the model with up to date data has achieved near 90% accuracy levels