US Accident Analysis
Machine Learning-powered traffic risk prediction platform designed to process and analyze over 7 million historical accident records. The system leverages Python, Pandas, and Scikit-Learn to perform large-scale data preprocessing, feature engineering, and predictive modeling, achieving 83.61% classification accuracy. The backend architecture consists of a FastAPI microservice responsible for serving a trained machine learning model through RESTful endpoints, enabling real-time traffic risk predictions. To support frontend consumption, the platform integrates a Java + Spring Boot Backend-for-Frontend (BFF) that securely orchestrates communication between services and delivers geospatial risk data to an interactive Leaflet-based map interface.
Phases:
[x] Phase 1: ETL, EDA, and ML Model Training
Cleaned and preprocessed 7 million historical records, conducted exploratory data analysis, performed feature engineering (including One-Hot Encoding for weather conditions), and trained a Scikit-Learn Random Forest model achieving 83.61% accuracy. Exported the trained model and feature structure into .pkl files for
[x] Phase 2: Data & AI Microservice with FastAPI
Built a high-performance Python server that loads serialized .pkl files into memory on startup for low-latency inference. Exposes a GET /data endpoint serving a 4,000-record sample dataset, a GET /columns endpoint sharing the model's feature schema, and a POST /predict endpoint for live inference on new accident data.
[ ] Phase 3: BFF (Backend for Frontend) Orchestrator with Spring Boot
Developing a strongly-typed API Gateway using Java and Spring Boot. Validates incoming UI requests with strict Jakarta schemas, handles data grouping and pagination, and acts as a secure reverse proxy to the Python microservice, eliminating browser CORS conflicts.
[ ] Phase 4: Interactive Dashboard & Simulator with React & Leaflet
Designing a Modern Dark Mode UI consuming structured data from the Spring Boot BFF. Features an analytical dashboard with interactive charts and geospatial visualizations powered by Leaflet, plus a dedicated 'Risk Simulator' form for live risk predictions from the ML pipeline.
Tech Stack
Python
Jupyter Notebook