Idea: Identify Potential
Future Missed Payments

r

Prediction of late credit card payments to aid Hyundai Card withprior to a missed payment, market repayment plans to high risk customers who may be struggling to make monthly paymentscapturing borrower changes in financial circumstances over timerepayment forecasting

Airflow
Training DAG

r

Develop I-tag that indicates borrow risk of a future missed credit card payment

a

Mock Borrower
Data

r

Develop prototype while awaiting permission to access Hyundai Card database

Hyundai Card
Borrower Data

r

Find similar tags to those in the prototype to implement a real prediction model

Load Task

r

Database-Prototype: Local Postgres ContainerDatabase-Production: Hyundai Card Hive DBPickle Pandas data frame and transfer its path to next task

Transform Task

r

Receive path from x-com and un-pickle data frameExecute Scikit-Learn transformation pipelinePickle transformed training and test data frames

Train Model 1

r

un-pickle training data frameimplement training function to produce modelpickle model

Train Model 2

Train Model N

Evaluate
& Select
Best Model

r

un-pickle training modelspredict test data labelsevaluate performanceselect and pickle best model for production

Jupyter Notebook

a

Exploratory
Data
Analysis

Transformation
Pipeline Development

r

Develop data transformation functions to be implemented by AirflowImplement tranformation functions in a transformation pipeline to be executable on both training and test data separately important for proper isolation during future training data refreshes (avoid training data getting into test data)

Database
Query

Model Development

r

Explore multiple algorithms and optimize with cross validation of model performance and grid/random search for optimal hyper parameters

Notify

Airflow
Prediction DAG

Load Unseen
Data Task

Transform Task

Prediction Task

Store Predictions
as I-Tags

I-Tags Represent
Risk to Future Missed
Payments

Pickle Best Model

Database Query

Database Insert

Unpickle Best Model