Development of machine learning algorithm for predicting sales of theatre tickets.

Purpose and goal

The purpose of the project was to utilize the extensive technical advances in machine learning and computing to help theaters and live entertainment producers predict how ticket sales will go during a season. Better forecasts can contribute to improved budgeting and more efficient marketing. The goal was to investigate the possibility of developing algorithms that can predict how the accumulated ticket sales will evolve.

Expected results and effects

The result of the project is that we built a comprehensive data-infrastructure to handle streaming ticket data and found a category of online learning algorithms that exhibit very promising features. The expected result is that we take these important building blocks to further develop this into a software system that can be used in production environment.

Planned approach and implementation

The project was divided into three phases. The first phase included literature studies and interviews. The goal was to identify what has already been done in this area and how intended users solve the problem today. Part two was about setting up the extensive data infrastructure needed to handle large amounts of streaming ticket transactions. Part three was about testing and evaluating different types of machine learning methods and models.