This project is to study 8950 credit card users and 18 credit card usage variables. The goal is to build a model for clustering the users for marketing strategy. The type of learning is unsupervised learning and the type of task is clustering.
With clustering, businesses can utilize their customer's data and group similar users into small groups for marketing purposes. It can help businesses reduce marketing costs and increase profit. It can also help businesses uncover the potential patterns in the data and open up new business opportunities.
I always get marketing emails about credit cards. But those promotions are not suited for me. In this dataset, I am interested in analyzing the relationship between credit limits, purchases frequency, cash advances, payments and balances. I aim to utilize the clustering technique to perform customer segmentation and create a marketing strategy.
The dataset is from a kaggle project named 'Credit Card Dataset for Clustering'.
Credit card dataset for clustering. (2018, March 2). Kaggle. https://www.kaggle.com/datasets/arjunbhasin2013/ccdata
The data set is a CSV format. This tabulated data has a data size of 8950 rows and 18 columns. The data types of the features are 14 float 64, 3 int64, and 1 object. The features are mainly related to balance, purchases, cash advances, credit limits and payments of credit cards.
- Load Data
- Data Cleaning
- EDA
- Feature Preprocessing
- Models
- Evaluation Models
- Hyperparameter Tuning
- Interpretation of the Clusters
- Conclusions