Project Outcomes
Investors on the Bridgin platform could not forecast or understand their expected cash inflows. Without visibility into anticipated repayments, investors could not plan effectively, assess portfolio health, or make informed decisions about future investment allocation. This gap reduced confidence in the platform and increased perceived risk for investors who relied on manual or incomplete tracking of repayment schedules.
This project set out to enhance Masunga’s Bridgin platform by developing features that improve investor decision-making and operational efficiency. The work focused on portfolio risk modeling, using historical contract and payment data to build predictive models that provide actionable insights into portfolio performance and repayment risk.
The project delivered a forecasting pipeline that shows Bridgin investors their expected monthly repayments per portfolio. It cut forecast error from roughly 46% on the baseline to 11%, and it separates error into repayment behavior and customer defaults so that misses can be explained. Clear data thresholds tell investors when a forecast is reliable and when to expect a range instead.
Key technology/tools used: Python(Segmentation, Decision Tree), Notion, AWS
Forecast approach
