Data Scientist- Loan Repayment (Model Built for predicting the loan risk and quality). Loan repayment model, idiosyncrasies for geographies, different bands of clearfraudscore partition

Guidelines (refer PDF for more detailed requirement):

1. Model Built for predicting quality on loan risk, and loan repayment for data in CSV (i.e. for Loan, Payment and Clarity Underwriting Variables), which are comma separated and column names in the first row.

2. Start with simple analyses as it may inform the model building processes;

3. It is open ended, for you to ask relevant questions about the data, do some preliminary exploration, perform the necessary manipulation or aggregations, generate visualization and reach conclusion and insights.

4. Thought process and ideas are integral! no right or wrong answers

5. Tips :
- Maybe the loan repayment value can be modified in order to make it easier to model;
- Perhaps there are some idiosyncrasies for geographies, or maybe different bands of clearfraudscore partition loan repayment nicely;
- Business context - what do you want to predict generally? Loans with what type of performance?
- Visualisation on insights and what do you understand from it;
- Data preparation, exploration and visualization;
- Clean coding style;
- The ability to tell a story using data;
- Clearly communicate your thought process;
- Back your assumptions with evidence.
- Modeling.



Deliverables :
1. Any languages of data analytic (Phyton or R is preferred, but SQL or others is still acceptable);
2. To write the commands, comments, or thoughts in IPhyton notebook in HTML format (word file is also acceptable, as long as readable) ;
3. Visualisation - any type (can be Tableau/ Google data studio or others as long as it is readable and can be downloaded)

Kindly submit the deliverables by end of this week (Friday - 11th November 2022) to me, and I will provide you a link to a Gdrive to drop the zipped folder.