Exploration and analysis of a dataset concerning the risk of credit default without any bank background, for the company HomeCredit.
My goal for this personal project was to explore and analyse the different features of the dataset to try and draw conclusions/parallels between a customer's credit default and its potential causal factors.
Once the analysis was made, I created an algorithm based on a Feed Forward Neural Network to classify and predict a potential credit default of a future customer, based on the relevant elements I found earlier in my EDA (Exploratory Data Analysis).
The content of this exploration can be viewed directly by browsing the Jupyter Notebook (.ipynb file).
For a matter of allowed space, the data used for analysis and prediction isn't in this repo. You can find it here
Here is my kaggle profile