Link to brief description from Day 1: https://docs.google.com/document/d/1WPKvRhCGiHHO837UbiYfESqJiFlrEhjFyPWArrh_MX4/edit?tab=t.0
Inspiration
I built StockSense because getting into the stock market felt incredibly intimidating, especially for beginners. It's often overwhelming to understand the data, figure out what risks are involved, or even know who to trust for advice. I realized people either spend hours researching themselves or constantly ask others for help to understand investments. My big idea was to create a friendly, all-in-one tool that automatically does a lot of that heavy lifting, making predictions, tracking, and learning about financial concepts clear and simple, so you don't have to do all the complicated work yourself or constantly reach out to others to figure out the risk.
What it does
StockSense is like your personal co-pilot for the stock market. It automatically predicts where a stock's price might go using smart AI, showing you trends on an easy-to-read chart, so you don't have to guess. You can also manage your own virtual portfolio and watch important stocks with live prices, seeing your investments' performance without manual calculations. My AI Financial Tutor powered by a Large Language Model (LLM) means you can just ask any finance question in plain English (like "What are the risks of penny stocks?"), and the LLM will understand and give you a straightforward answer. This means you don't have to dig through endless articles or ask friends to understand complex topics or potential dangers.
How we built it
I created StockSense using the Python and a tool called Streamlit, which let me build the actual web app. I used libraries like yfinance to automatically grab all the live and historical stock data, and TensorFlow to build the AI that forecasts prices. The Large Language Model (LLM) was mainly used to make the finance and statistics easy to understand for beginners. This LLM is designed to understand your questions about the market and generate clear explanations, effectively acting as an intelligent assistant that helps you quickly grasp concepts and potential risks without needing to be an expert yourself or consult multiple sources.
Challenges we ran into
A big challenge was how slow the app was when first loading, especially with all the stock data it needed to return and the AI prediction model to prepare. I had to go back and change the code many, times, trying different ways to optimize things and speed up the loading process significantly. Also, this was my first time using a Large Language Model (LLM), so figuring out how to make an AI truly understand and simplify complex financial information to help users assess risk was a big challenge.
Accomplishments that we're proud of
The thing I'm most proud of is the AI Financial Tutor, powered by an LLM. It took a while for me to understand how it worked, so learning and successfully applying that to my project was one of my biggest accomplishments.
What we learned
I learned how complicated stock market data can be and how to use AI like LSTMs to find patterns in it. I also became much better at building interactive web apps with Streamlit. Most importantly, I really saw the incredible power of Large Language Models (LLMs) firsthand, as this was my first experience working with them.
What's next for StockSense
My biggest goal is to fully integrate the LLM so it can do even more. It being able to read all the latest financial news and company reports, and then summarizing potential risks, opportunities, or sentiment for you automatically, so you don't have to do that complex analysis yourself or rely on others' opinions is my main goal. I also want to add more simple tools for beginners, improve my prediction AI even further, and make the app easier to use on any device.
Log in or sign up for Devpost to join the conversation.