Money Translate

Inspiration

As students who are just now starting to manage our own personal finances, we find that the amount of technical expertise that is often required to understand these documents can be extremely overwhelming. Furthermore, it creates barriers to entry in being able to "explore" and learn about finances and investments.

When researching the amazing sponsors for UofT Hacks X, we were especially intrigued by both The Advice Company and Cohere AI. As individuals who love reading about the financial world, we found The Advice Company prize category to be a great fit for our interests. Moreover, we felt that Cohere AI's natural language processing model was such an incredible technology that we wanted to incorporate and learn to use in a meaningful way.

What it does

We developed a website called Money Translate. This website allows user's to input a PDF file of a financial document and it will then output a summary that details the most important concepts of the document in a manner that is easy to understand.

How we built it

We built the frontend of the website using React, HTML, CSS, and JavaScript. We built the backend of the website using the Python-based Django webframework and we achieved the text summarization feature using the Cohere API. In addition, we hosted the project on the website PythonAnywhere.

Challenges we ran into

One of the main challenges we ran into while developing this project was that the text summarization feature of the Cohere AI was not providing the results we intended it to, despite various attempts at troubleshooting the issue.

Accomplishments that we're proud of

This was Vivek Jariwala and Jeffano John's very first experience in developing a frontend using React. While the learning curve was initially quite high, we are very proud with the design of the frontend of the project and the user interface principles that we incorporated.

In addition, Nigel Lobo and Michael Gao learned how to properly configure Django in order to serve the frontend that their teammates had developed.

What we learned

Vivek Jariwala and Jeffano John learned how to create a website using React, user interface principles, and what makes an attractive design for websites.

Nigel Lobo talked to the Cohere AI employee at the sponsor booths section of the hackathon to learn more about how to manipulate the data the was received by the API.

Michael Gao learned how to use React and how to utilize the Cohere API.

What's next for Money Translate

Due to the time constraints of the hackathon and difficult schedules of being full-time university students, our group was unfortunately not able to complete the project. While both the backend and frontend were finished, they did have enough time to integrate them together.

In terms of features that we would expand on, we would like to improve the quality of Money Translate's text summarization.