We want to make it easier for the millennials to save money and reinforcing good money spending habits but also put an emphasis on money security with machine learning techniques.
What it does
Our service is used for automatically taking a small amount out money with each transaction a user makes with their current account and automatically puts the money into his saving account which they can later use.
How we built it
Our backend was written in Python. We use isolation forest to detect fraudulent transactions. Flask was our framework of choice. We integrated our backend system with Kafka as a proof of concept. On the frontend, we use VueJS to visualise the transaction list and create dummy transactions that go through the complete workflow(transactions are piped into Kafka and later consumed & charges calculated / evaluated fro fraud).
Challenges we ran into
- Fraud detection is hard
- Thinking of a way to distribute the system
Accomplishments that we're proud of
- We built a system to detect fraudulent transactions
What we learned
- Fraud detection is a really broad and hard topic
- Kafka is easy to setup and can run 'out of the box'