How mrbeastgamerzz/credit-risk-delinquency-analytics_project works
This is a credit delinquency analytics project that turns customer records into risk scores, explanations, and collections dashboards. The documented workflow moves from Excel exploration and cleaning through SQL analysis, Python preprocessing and modeling, SHAP explanations, and score export to Power BI. The Python excerpts show individual transformations and exports, but not a runtime orchestrator linking the scripts; phase-to-phase edges below are therefore based on README documentation. Excel, SQL, and Power BI internals are represented only at the level documented in the README and tree.
Main entry points: featured.csv Featured Data · exporter.py Score Export · geldium.db SQLite Risk Data · Cleaned_dataset.csv Excel Cleaning
Components 20
Excel Analysis
- Raw Customer Data
excel data/raw/Delinquency_prediction_dataset.xlsx - Excel EDA
excel data/eda in excel/Delinquency_prediction_dataset-eda.xlsx - Excel Cleaning
excel data/processed/Cleaned_dataset.csv
SQL Risk Analysis
- Portfolio Queries
sql/SQL_queries.sql - SQLite Risk Data
sql/geldium.db
Python Analytics
- Data Preprocessing
src/preprocessor.py - Cleaned Data
excel data/fully_cleaned.csv - Feature Engineering
src/feature_engineer.py - Featured Data
excel data/featured.csv - EDA Visuals
src/eda_plots.py - Model Training
src/model.py - XGBoost Model
outputs/model/xgb_model.pkl - Evaluation Metrics
outputs/reports/metrics.json - SHAP Explanations
src/explainer.py - Customer Risk Factors
outputs/reports/shap_summary.csv - Score Export
src/exporter.py
Risk Delivery
- Dashboard Dataset
outputs/reports/dashboard_data.csv - Collections Dashboard
powerbi/risk_engine_dashboards.pbix
Other components
- Risk Team
- Collections Stakeholders
Connections 24
- Raw Customer Data → Excel EDA: explores
README.md - Excel EDA → Excel Cleaning: cleans
README.md - Excel Cleaning → SQLite Risk Data: loads data
README.md - Portfolio Queries → SQLite Risk Data: queries
README.md - SQLite Risk Data → Collections Dashboard: feeds foundation
README.md - Risk Team → Excel EDA: explores portfolio
README.md - Excel Cleaning → Data Preprocessing: provides input
README.md - Data Preprocessing → Cleaned Data: writes
src/preprocessor.py - Cleaned Data → Feature Engineering: provides input
README.md - Feature Engineering → Featured Data: writes
src/feature_engineer.py - Featured Data → EDA Visuals: provides input
README.md - Featured Data → Model Training: trains on
README.md - Model Training → XGBoost Model: saves
src/model.py - Model Training → Evaluation Metrics: writes
src/model.py - Featured Data → SHAP Explanations: explains predictions
README.md - SHAP Explanations → XGBoost Model: loads
src/explainer.py - SHAP Explanations → Customer Risk Factors: exports factors
src/explainer.py - Featured Data → Score Export: scores customers
README.md - Score Export → XGBoost Model: loads
src/exporter.py - Score Export → Customer Risk Factors: merges factors
src/exporter.py - Score Export → SQLite Risk Data: writes scores
src/exporter.py - Score Export → Dashboard Dataset: exports
src/exporter.py - Dashboard Dataset → Collections Dashboard: powers
README.md - Collections Dashboard → Collections Stakeholders: presents risk
README.md