Credit Card Data Clustering Analysis using Kmeans on BangDB

Problem Statement

This case requires developing a customer segmentation to define marketing strategy. The Dataset summarizes the usage behavior of about 14360 active credit card holders during the last 6 months. The file is at a customer level with 18 behavioural variables.

This is taken from Kaggle, here is the problem statement in detail.


Build a model to perform customer segmentation.

Get help doc

Download the file and follow the help doc to complete the task.


Data description

The dataset contains 14360 events and 18 attributes. The dataset consists of both continuous as well as categorical attributes.


The given problem is a regression problem. To solve it, we will be using the Support Vector Regression algorithm (SVR). The advantage of SVR over linear regression is that SVR gives us the flexibility to define how much error is acceptable in our model and will find an appropriate line (or hyperplane in higher dimensions) to fit the data.


There are three ways to train a model on bangdb. One is to directly register the meta_data for training (we call it json request which contains all the details about the model), the second is to create mage_data for training by following the workflow on CLI and the third is by using APIs.

Here we will be training the model using all methods. On bangdb, we have the option of training a model from a file( file format can be CSV or JSON) containing training data or from streaming data. Here we are training the model from a CSV file.


Using CLI workflow.