SimpliFeed

SimpliFeed is the ultimate financial literacy tool for younger audiences. Our NLP algorithms simplify financial articles, making them easier to understand and more enjoyable to read, suitable for all.

Comment

Problem Statement

How can ML services help consumers make more informed financial decisions?

Inspiration

The inspiration and rationale behind SimpliFeed is to help young investors like us better obtain financial information through news articles by coming up with a solution for the common problems that young investors face when trying to stay informed about the financial markets. As young investors, we often struggle with information overload, complex jargon and a lack of context when reading financial news articles. Hence, we hope to come up with an app that can easily identify the most important and relevant points in news articles without having to wade through a large amount of irrelevant or extraneous information.

What it does

SimpliFeed is a fintech-as-a-service platform that uses natural language processing (NLP) to simplify financial articles and make them more easily understandable for a younger audience. With SimpliFeed, users can access a wide range of financial articles through our website, and can easily toggle between the original version and a simplified version created by our NLP algorithms. The platform also offers personalised recommendations based on users' interests and financial goals, helping them to make informed financial decisions. With its combination of NLP-powered simplicity and interactive learning tools, SimpliFeed is a powerful resource for helping younger audiences understand finance and make smart financial decisions.

How we built it

Our frontend was written in Vue.js while our backend makes use of python and FastAPI. Additionally, our app interfaces with a custom-trained machine learning model deployed using Hugging Face, AWS Sagemaker, AWS Lambda and AWS API Gateway.

Challenges we ran into

We initially made use of the HuggingFace Inference API to deploy our ML model. However we eventually were faced with response time issues that affected the user experience. We ended up making use of AWS to solve these issues. Additionally, scraping the internet for relevant financial news also proved difficult due to the variety of news sources available, and the ability for websites to detect automated website accesses. Some clever scripting in Selenium helped us to get work around this problem.

Accomplishments that we're proud of

  • Clean and appealing user interface
  • A reliable and accurate ML model

What we learned

  • ML Model Deployment
  • Search Strategies in ML models (i.e Beam Search, Greedy Search, etc ...)
  • HuggingFace API
  • AWS Deployment / API

What's next for SimpliFeed

  • A slider to allow users to change the length of the summarised text according to their preference
  • Adding more features to faciliate learning and possibly offer personalised recommendations and interactive learning tools

Built With

Share this project:

Updates