This is a full end-to-end data analytics project simulating a loan application funnel at a fintech or lending company. It includes:
✅ Funnel performance monitoring
✅ Automated KPI reporting and alerting
✅ Advanced data insights and cohort analysis
✅ A/B testing to evaluate underwriting strategy changes
✅ A deployed interactive dashboard (via Streamlit)
To simulate, analyze, and optimize the loan funnel journey — from application to funding — and identify areas of improvement using metrics, insights, and A/B experimentation.
| Category | Description |
|---|---|
| 🔁 Synthetic Data Generation | Realistic data for 10,000 applicants with credit score, income, loan amounts, funnel stages, approval, and default outcomes |
| 📊 Funnel Analysis | Stage-wise conversion rates, approval and funding rates, weekly application trends |
| 📈 Advanced Insights | Analyze how age, income, and credit score influence approvals. Cohort analysis by credit bands and income brackets |
| 🧪 A/B Testing | Compare approval and default rates for different underwriting strategies. Perform Z-tests for statistical significance |
| 🚨 Automated Reporting | Scheduled KPI monitoring and alerting if metrics fall below defined thresholds |
| 📊 Interactive Dashboard | Deployed with Streamlit to visualize KPIs, test results, cohort breakdowns, and alerts |
- Python, Pandas, SQLite
- Faker for synthetic data generation
- Statsmodels for A/B testing (Z-test for proportions)
- Streamlit for dashboard deployment
- SQL for query-based analysis (via SQLite)
- Clone the repo
- Install dependencies:
pip install -r requirements.txt- Generate data and SQLite DB:
python src/generate_data.py- Launch the dashboard:
streamlit run dashboard/app.py