Partnering with RPI graduate students to explore customer projection models
Over the past semester, we’ve worked with students studying business analytics at Rensselaer Polytechnic Institute (RPI) to build customer projections. Genadi Naydenov (my fellow data-focused balenista) and I partnered with RPI’s Lally School of Management in their Supply Chain Capstone course. Professor T. Ravichandran, teaches the course and mentored the student team in this project.
Congratulations to the team
The project was done by a team of graduate students in RPI’s Business Analytics Program. The team members were:
balena’s mission is to provide infrastructure for fleet owners so they can focus on developing their applications and growing their fleets with as little friction as possible.
As a company, we focus on long time horizons, emergent complexity, and feedback loops. This means it is both difficult and important to understand the impact of strategic decisions. For that purpose, we are developing a simulation framework that models users’ journeys from the bottom up. It simulates each user from their first discovery of balena, through their signup and fleet growth, and to their conversion to paying customer.
The goal of the collaboration was to forecast customer device growth and churn models based on historical data. The RPI team built two models (Stochastic Gradient Descent and Autoregressive Integrated Moving Average) to predict device growth and one model (Cox Proportional Model) to analyze customer churn, based on an anonymized sample of our customer data.