Explore and model bank customers behaviour with classification algorithms

Project • Data source • How To Use • Process

The project

The aim of this (learning) project is to analyse the characteristics and behaviours of bank customers in order to:

• predict and limit the churn rate
• create clusters of resembling customers to design targeted communication campaigns

A dashboard accessible to all interested people in the company will showcase the findings.

Data source

The dataset can be downloaded on the following link: Credit Card Customers.

How to use

For end users, just go to the dashboard online. Here is a short video of how it looks like. Enjoy!

For developers, you'll need Python installed on your computer to clone and run this application. From your command line:

# Clone this repository
$ git clone https://github.com/lyesds/churn_prediction

# Go into the repository
$ cd churn_prediction

# Install dependencies
$ pip install requirements.txt

# Run the streamlit app
$ streamlit run dashboard.py

Process

Explore

• Univariate
• Bivariate

Predict churn rate

• Classification machine learning algorithm

Build clusters

• Clustering

---

GitHub @lyesds  ·  LinkedIn @lyes