Mobile & Data
Credit Risk Analysis
XGBoost predictive model for financial risk classification
A comprehensive credit risk classification project built during a virtual internship with id/x partners x Rakamin Academy. Used XGBoost and advanced feature engineering to predict credit default risk, with full exploratory data analysis and stakeholder-ready visualizations.
Workflow
- 1Exploratory data analysis (EDA) to identify patterns and outliers
- 2Feature engineering and missing value treatment
- 3Baseline model comparison across multiple algorithms
- 4XGBoost model tuning with cross-validation
- 5SHAP-based feature importance analysis
- 6Presentation deck with data storytelling for stakeholders
Impact
Achieved high classification accuracy for credit default prediction. The project demonstrated production-ready data science skills and earned completion recognition from id/x partners.
Key Features
- Full EDA with statistical analysis
- XGBoost with hyperparameter tuning
- SHAP feature importance visualization
- Stakeholder presentation with insights
Tech Stack
- Python
- XGBoost
- Pandas
- Matplotlib
- SHAP
- Scikit-learn