As per a study conducted by CGI, banking customers of leading banks want banks to help them save money.

  • 52% of the customers said, “Tell me what I am spending money on and how I can save”.
  • 55% of them wanted access to “wealth building advice by leveraging their financial data”.

Today, few years later this study, these demands are becoming a reality. In this information age, customers of banks want their bank to evolve with their changing needs and offer more value in return of their relationship with bank.

We want to offer a solution on the FFDC platform that addresses both the above needs.

What it does

Our solution intends to help:

  • Banks, by being able to provide more value to their customers as an account holder by automatically offering new products that suite particular financial behavior of a customer.
  • Customers, by being able to get valuable and easy to understand analytical insights into their financial lives and receive wealth creating recommendations based on their financial behavior.

So, the idea is to automatically review account holders spending patterns and suggest an appropriate bank product that will help account holder make better investment choices.

Our Mission:

  • The new app will help account holders manage expenses by showing statistical data about their spending patterns, using machine learning.
  • The idea is to show account holders how their monthly cash-flow is made of, and predict of their account balances in the future if they continue with current spending patterns. This is particularly useful for customers with varying income sources. For e.g. Small-Medium-Enterprises, Freelancers, Artists, etc.
  • Find a match with a bank product and make suggestions. Show how this product impacts there future bank balances using.
  • All this with a beautiful visual representation of the data that the users will fall in love and feel engaged to use it.

A lot of such efforts have been made in the stock trading platforms, but the banking sector seems to have fallen behind to catch-up. So we see good use-case for this solution in the future.

How I built it

We're a team of enthusiastic developers/QA and figured out that Python would be a quick way to start working on a solution like this. We used Pandas library for data parsing for bank account statements and were able to perform analytics on it. We pulled in interesting insights from this info for e.g. find recurring patterns of expenses, see how the cash-flow is made of, what are the spending patterns etc. We wanted to show this on a nice to look at GUI so that they look different or refreshing when compared to a typical bank account statements. We used Dash Graphing framework from Plotly to build the Graphs and GUI.

We used FFDC Get transaction details APIs to pull data from FFDC. But since not many records were available on FFDC we continued to use the data set of transactions we built for the test purpose.

We also added a Upload your own Account Statements feature, so that any user not on FFDC can also chec out the app.

Tech Used

  • Python for Backend
  • Pandas for Dataprocessing
  • Dash by for Graphing framework
  • APIs to get transactions data
  • Naive Bayes ML for Automatically Categorizing transactions into specific categories like Groceries, Fuel, Dining out, etc.

Challenges I ran into

All the technologies we used here were very new. Python and Dash was completely new and everyone had to go through a very-very-short learning curve to learn and use the tech. We had to plan well in advance and split-up tasks in the team appropriately. All this with balancing the work in our regular work was a challenge.

This challenge of learning and working on the new technologies than the ones we actually work on, or have expertise on in our office projects was huge. So, the moments we completed our working prototype, we had already won our own Hackathon :)

Accomplishments that I'm proud of

We're proud of the team work we were able to pull-off. Even though everyone was new to the tech we used we built up a nice app that was liked by our colleagues. We're also proud of the enthusiasm which helped us make this big journey of change of a mindset and learning/implementing on a different kind of problem which none of us had worked on before. In the end it was great learning experience for everyone on the team and we are glad of it.

What I learned

A positively spirited attitude can help you cross any difficulties that may come. In our case, everyone worked on a different portion of the app by learning himself/herself and then helping the other person do the same. That was great collaboration.

What's next for Fusion Wealthify

Currently we're using Machine learning to automatically categorize the transaction based on our pre-trained data set. We want to complete the app for the remaining portions and add better machine learning capabilities based on huge Datasets of data that can only be available with a bank. Find more insights that would be beneficial for all categories of accounts holders.

Built With

Share this project: