This project was part of the hiring process for a Data Scientist position @ Credix.

The proposed was to build an analytical strategy towards predicting time to payment in a portfolio of Brazilian companies having at disposal assets data, company information and credit bureau data.

For this project mostly exploratory work was done on Jupyter Notebooks (folder notebooks) ordered from 1 to 4. Also, some helper functions were developed throughout the developed and shared in "case/helper_functions.py" package.

AS an effort for this exercise I have developed a :

• duration table for assets data
• data enrichment including company and bureau data
• specific analysis of large clients of Credix
• Attempted to model time to payment using AFT Weibull models

I was not able due to the deadline to draw a deploy strategy for this product, but I would appreciate any feedback on my work so far.

For running on this repository, please have poetry package manager in your machine and install all dependencies in the toml file.

Have a great day ;)