Churn Project

The idea of this project is simple, detect bank fraud, to do this I am using a dataset called "Bank Account Fraud Dataset Suite (NeurIPS 2022)".

In the beginning of this project I chose to go with an autoencoder this is because of my hardware limitation using an AMD GPU and the need to use Tensorflow-directML.

Left Image

Using the graph from the data below it seems to indicate that higher the social score higher retention of customer

This brings up many questions such as:

• Low Engagement results in high churn, low social score indicates a higher chance at churning, less engaged with the service? service issues?

• Social influence appears to matter in terms of churn.

Actions:

• Target Low-score - Personalized offers?

• Leverage the High scorers - REWARD them for loyalty, offers only attainable through loyalty, more referrals mor rewards.

IBM Telco Customer Churn dataset (modified) #1

Links

Original Dataset license

JB Link Telco Customer Churn
https://www.kaggle.com/datasets/johnflag/jb-link-telco-customer-churn

Authors

Joao Bandeira, Jack Chang