Case Study: Zippi improves credit risk assessment and data normalization with Hevo Data

Zippi Fuels Financial Inclusion With Hevo By Streamlining Data For Accurate Risk Assessment

Zippi, a Brazilian fintech company providing working capital to small entrepreneurs, faced significant challenges with data integrity and normalization for its machine learning risk analysis models. Its data was siloed across sources like Typeform and Amazon S3, and the manual process of normalizing complex JSON files was slow and inefficient. This fragmented data landscape, managed with a previous platform (DMS), hampered model performance and raised security concerns, prompting the search for a new solution from Hevo Data.

Hevo Data implemented its zero-maintenance data pipeline platform to automate ingestion and transformation. The solution provided a unified data platform, streamlined JSON normalization, and enabled hybrid schema mapping, which drastically reduced the process from days to just minutes. For Zippi, Hevo’s solution resulted in improved data quality for its ML models, a measurable reduction in loan default rates, and enhanced productivity for its data science team, allowing for faster and more accurate creditworthiness assessments.

Zippi

Bianca Bartolomei

Data Engineer