Projects

Python Predicting credit card customer churning and devising strategies for customer retention

Kaggle version

Image source: debt.com

• Used CatBoost to predict whether a customer will churn based on his/her demographic information and transaction amounts and counts with the credit cards of the bank. The model achieves an F1 score of 0.911 and identifying the total transaction amounts and counts in the last 12 months as the most important features in predicting customer churning

Python Predicting customer purchase of travel insurance with XGBoost

Image source: Cover Karo

• Built an XGBoost classification model with 84.3% accuracy in predicting whether a customer will purchase travel insurance based on factors like age, annual income and travel history etc

R When ICTs merge with governance in the 21st Century: an investigation of the association between e-governance advancement and governance quality of UN member states from 2003 to 2018

Image source: Go News India

• Tested hypotheses about whether advancement in ICTs would improve the governance quality of a country by using mixed effects model to analyse time series data compiled from the UN’s E-Government Development Index and the QoG dataset