Thinking about how many customers companies loose, especially in subscription services, we became fascinated by the possibilities and ways in which beautiful software could solve the problem.

What it does

Our project uses artificial neural networks to predict the possibility of a person exiting a subscription service (in this case - a bank) thus helping the service provider find possible schemes/strategies to prevent the above from happening.

How we built it

The actual model is built using python’s machine learning libraries (tensorflow, Keras), the frontend is made using html and css, and the backend was built using flask.

Challenges we ran into

One of our biggest challenges was to integrate the model with the backend and actually bring up the accuracy using hyper-parameter tuning.

Accomplishments that we're proud of

One of the accomplishments we’re most proud of is actually having a working model. We are also proud of the way we worked in a team even though we came from different schools and did not know each other from before.

What we learned

We learned a lot about how different technologies work together to deliver a final product. Although one person on the team knew data science, one knew backend, and the other knew frontend, it was hard to bring all the three aspects together but was a really good learning experience.

What's next for is looking forward to not only work in the financial sector, but also freemium model subscriptions and actually running dry runs for startups to possibly change strategies by collecting and working on better datasets.

Built With

Share this project: