Standard Bank Tech Impact Challenge: Xente credit scoring challenge
2000 Zindi Points
Can you predict the likelihood of credit default of ecommerce clients?
367 data scientists enrolled, 96 on the leaderboard
30 August 2019—2 December 2019
Data exploration, trying to understand the problem and feature ideas Notebook.
published 5 Nov 2019, 17:15
edited 14 days later

I've just started working on this data, here is my repo for this challenge where i'm trying to understand the data and the problem to solve. I will be updating this notebook as soon as i make some progress.

https://github.com/blenzus/StandardBankLoanDefault/blob/master/Standard_Bank_EDA.ipynb

https://github.com/blenzus/StandardBankLoanDefault/blob/master/Standar_Bank_BaselineModel.ipynb

Update : Exploration is somewhat well developed. Posted a starter notebook in Python ( Jupyter Notebook ) that gets you around the 30th spot. Intuition behind the features generated in the baseline model is in the other notebook in order to understand the logic behind such features.

EDIT : The features in the starter notebook can get you up to 0.8 Auc ( 2nd spot ). Use them wisely.