Detection of suspicious loans using clustering techniques. Performed preprocessing, normalization, and analysis of financial data to identify risk patterns using K-Means clustering. The optimal number of clusters was determined with a Silhouette score of 0.91 and a Davies-Bouldin index of 0.90. Results were further analyzed using PCA for dimensionality reduction. Tools: Python, Pandas, Matplotlib, Scikit-learn.

This project was developed in collaboration with Daniel Roldan Serrano (https://github.com/danirold)