How sahbimlesh007/credit-risk-scoring-system works
This is a browser-based credit-risk scoring application. An applicant enters loan and credit-history details, the browser submits them to the FastAPI prediction endpoint, and the service preprocesses the application, scores it with a loaded CatBoost model, applies a decision threshold, and returns risk and confidence metrics for display. The browser also supports exporting a report. The sampled source confirms the request, scoring, and display flow; training-data and notebook roles are not shown, so they are omitted.
Main entry points: main.py Prediction endpoint · script.js Form and results · index.html Application form · CatBOost.pkl CatBoost model
Components 10
Web interface
- Application form
static/index.html - Form and results
static/script.js - Risk assessment
static/script.js - Report export
static/script.js
Prediction service
- Prediction endpoint
main.py - Input preprocessing
main.py - Risk scoring
main.py
Scoring assets
- CatBoost model
CatBOost.pkl - Decision threshold
catThresold.pkl
Other components
- Applicant
Connections 11
- Applicant → Application form: enters application
- Application form → Form and results: submits form
- Form and results → Prediction endpoint: POST /predict
- Prediction endpoint → Input preprocessing: validates and prepares
- Prediction endpoint → Risk scoring: requests prediction
- Risk scoring → CatBoost model: predicts probability
- Risk scoring → Decision threshold: reads threshold
- Risk scoring → Prediction endpoint: returns assessment
- Prediction endpoint → Form and results: returns JSON
- Form and results → Risk assessment: renders results
- Risk assessment → Report export: exports report