EQA AI

EQA AI is a new AI powered investing platform that makes investing easier through leveraging LLM's to process vast amounts of financial data in an easy to interpret way for investors to make decisions

Comment

Inspiration

I was inspired to make this because as an investor the more I began to learn about investing, the more I realized just how disadvantaged the every day Joe is in the stock market. There is so much data out there needs to be considered when making accurate investing decisions that one person simply cannot comprehend. Hence I decided to make EQA AI to create something investors can use to be able to analyze these vast amounts of data to be able to compete with the institutional investing corporations who have historically had access to this data and have used it to manipulate the market for centuries. EQA AI aims to equal the playing field of investing and help the average person succeed in their investing.

What it does

It takes in various financial data like Company Financials, Market Sentiment, Sector News, and Insider Trading data this is accessed live and that data is then sent to Claude to analyze it and then return a number that indicates the confidence in the stock through its objective analysis of the data form 1-10 with 1 being low, 5 being average, and 10 being high. This allows the users to get a nice analysis of the data in a way they can understand very easily and from there they can decide wether to buy, hold, or sell certain stocks.

How we built it

I built it by first accessing all the financial data through Finhub's API to get all the data and aggregate in a json. Then that data is sent to Claude with the appropriate system prompt to perform the analysis and return it in a very easy to understand format for the user. The AI portion of the analysis was done through fine tuning a Claude prompt to get it to understand the large amount of financial data I was sending to it and most importantly not only compute it but also be able to explain why it made its analysis that way to minimize possible hallucinations.

Challenges we ran into

A challenge I ran into was getting the data from claude to show up in the format I wanted since there were many problems getting the data to Claude and getting the data back from it. I am not good at the terminal and this project required a lot of terminal interaction to install packages and test servers as well as debug so that was certainly challenging.

Accomplishments that we're proud of

I am very proud of our homepage since it looks very sleek and modern and took a lot of time to make. As well as how smooth everything flows in this project. This was the first time I've ever had a website look anything remotely as good as this.

What we learned

How to use python to handle API's and data and integrate it with HTML websites. Also how to create cool animations and background for websites. This was also the first I ever got to interact with any AI API, so I learned a lot about the importance of ensuring the data you are sending and receiving back matches the type that you were expecting.

What's next for EQA AI

I will continue expanding EQA AI to make choosing and selecting the stocks easier and adding graph's and other things to make the analysis less text based and more visual based. Right now its got a lot of text so if I can find a way to add some nice visuals that would make it a lot easier for the users. Another thing would be making it so that there is a nice dashboard of all the trending stocks so people don't need to manually enter the stock and sector and can just get a nice visual way of selecting which stock to analyze. Down the line, this could scalable by having this stored in a database so if 2 people run the analysis at similar times it can check if an analysis was already done before and just use that data instead of having to recompute.

Built With

Share this project:

Updates