The customer and transaction data used can be collected from the following link below
https://www.kaggle.com/shivamb/bank-customer-segmentation/
If the dataset and/or link no longer exists, please contact me at [email protected] and I will send it to you accordingly.
The data consists of transactional, demographic, and geographical information for bank customers in an Indian bank.
In a high level, the project consists of the following steps:
- Exploratory data analyses to uncover customer behavior and demographic trends.
- Features were created to enrich segmentation analysis.
- Data was cleaned and prepared for the KMeans clustering model.
- Hyperparameter tuning to identify optimal number of clusters.
- Post hoc analysis to identify cluster specific marketing as well as customer acquisition recommendations.
The "Customer Segmentation Project.ipynb" file was initially run in Python 3.8.8. However, Python 3.8 is already obsolete when I checked on April 9, 2025. So we will use Python 3.10. setup.sh downloads the necessary Python packages, gets a virtual environment set up, and runs the jupyter notebook. requirements.txt consists all the Python packages that are needed in this project. The instructions are provided below.
- Use a linux Ubuntu environment
- Run setup.sh
- In the terminal, click on the link provided. You will see the Jupyter notebook UI in port 8888.
You should now be able to run the jupyter notebook