Customer Segmentation in the US
In this project, we’ll use data from the 2019 Survey of Consumer Finances. First, you’ll identify households that have a hard time getting credit. Then we’ll build a model to segment these households into subgroups. Finally, you’ll create an interactive web app to share our work.
This project is an example of unsupervised learning, specifically clustering. It can be used in commercial contexts for marketing or customer segmentation or in sociological contexts to study social stratification.
We were able to:
- Compare characteristics across subgroups using a side-by-side bar chart.
- Build a k-means clustering model.
- Conduct feature selection for clustering based on variance.
- Reduce high-dimensional data using principal component analysis (PCA).
- Design, build and deploy a Dash web application.
This project consists of 5 Notebooks:
- Exploring the Data
- Clustering with Two Features
- Clustering with Multiple Features
- Interactive Dashboard
- Small Business Owners in the United States🇺🇸
You can find the whole project blog here