Dataset: Credit Card Customer Dataset
The dataset shows the cutomer's spending behaviour on various credit cards.
Business Problem - The bank would like to know the distinct groups of customers existing in their bank. And, with the outcome of the analysis the bank would provide marketing offers to relevant customers.
Steps for python implementation
- Read and select numerical feature for clustering.
- Outlier treatment and impute missing values
- Feature Engineering
- Perform standardization of features to remove impact of units and enable distance computation
- Remove correlations
- plot WCSS graph(elbow curve)
- Build clusters with K-Means and Hierarchial clustering
- Evaluate clusters and build profiles to reveal patterns(cluster profiling)