Understanding Credit Insecurity: A Data-Driven Segmentation of American Households (Using the 2022 Survey of Consumer Finances)

Every three years, the U.S. Federal Reserve conducts the Survey of Consumer Finances (SCF) to collect detailed information on the financial well-being of households across the country. In the 2022 edition, thousands of data points were gathered, ranging from income levels and debt obligations to demographic profiles and financial behavior. In this project, that data becomes the basis for a practical exploration. The task begins with identifying households that report difficulty accessing credit - a critical issue with implications for economic participation. From there, the goal shifts to organizing these households into distinct groups based on shared financial and demographic characteristics.

To achieve this, the project applies an unsupervised machine learning technique known as K-Means Clustering. The data is first prepared through feature selection and dimensionality reduction using Principal Component Analysis (PCA), and then segmented using the k-means algorithm. The output is a set of well-defined clusters that reveal patterns within the credit-constrained population. To ensure the results are accessible and usable, the final step is the development of a web-based dashboard using Plotly Dash. This web application allows users to interact with the data, compare clusters, and derive insights that can inform further analysis or decision-making in fields like marketing, public policy, or financial services.

---

Watch the Demo Video!

Access the Colab Notebook