Creator / project
Built to prove the full machine-learning lifecycle.
Flowcast is an end-to-end urban forecasting project by Davis Higgins, a Data Science student at UNC Charlotte, Data Analyst, software/web developer, and founder of Higgins Digital. The project connects public data engineering, predictive modeling, validation, AWS MLOps, API serving, geospatial visualization, and product design in one working system.
Why Flowcast
A model is more useful when it can survive outside a notebook. Flowcast was built to carry a forecasting problem from raw public data through feature engineering and evaluation, deploy it through AWS, and make the result understandable through a real interactive product.
- Public data engineering
- Leakage-safe feature engineering
- Global XGBoost modeling
- Time-based validation
- Quality-gated model registry
- Serverless inference API
- Geospatial product design