Member_Retention (Monkey)

We've utilized machine learning to predict which Servus members are likely to close their accounts and then displayed the information for service representatives to provide care to customers in need.

Comment

Inspiration

Customer retention is a focal point for many businesses. The most valuable customers are the loyal ones, especially in banking and finance.

What it does

We've developed a machine learning model that looks at a customer's metrics including age, income, number of accounts, credit score, and more.

The web-app interface will provide a service representative with client information, and an assessment of their likeliness to leave the credit union.

How we built it

Out project consists of a dataset generator written in Go, Python for ML, and Django + bootstrap for the web-app.

Challenges we ran into

The first challenge was generating good data as we were lacking good real data given its personal nature. The model uses various assumptions that would have to be tweaked for real-world data.

Accomplishments that we're proud of

  • Generation of a semi-realistic dataset
  • Implemented shallow supervised ML models
  • Used Django built-in user authentication.
  • Successfully wrote a script that imports the generated .csv data into a Django model.

What we learned

None of us used Django really before this hackathon, so we're proud to have tried a new tool.

What's next for Customer_Retention (Monkeys)

Industry experts could help tweak the model and it could be applied against real-world data to help Servus Credit Union.

Share this project:

Updates