End-to-end address-level scam risk modeling on Ethereum: build behavioral features from transactions, train a supervised model, and test how well it holds up under time drift and dataset shift (including an external regulator list: CA DFPI).
- Notebook Dashboard: https://github.com/LittleHouse75/flatiron-ml-modeling-pipeline/blob/main/notebooks/00_Overview.ipynb
- Pitch: https://github.com/LittleHouse75/flatiron-ml-modeling-pipeline/blob/main/flatiron-ml-modeling-pitch.md
- Reflection: https://github.com/LittleHouse75/flatiron-ml-modeling-pipeline/blob/main/reflection.md
- Video Presentation: https://youtu.be/31tma5EMXwE
notebooks/— EDA, feature engineering, modeling, and evaluations (see dashboard link above)src/— feature engineering + reusable pipeline code (ex:src/featureeng.py)data/— expected inputs / generated artifacts (including DFPI-related files, if present)