Inspiration
We like money and now we're gonna use this app to make money.
What it does
We've created a stock price prediction app that takes natural language input such as a company name, a ticker symbol, or a general request. It uses Gemini's API to convert the input into a valid stock ticker symbol or returns an error if the input is invalid. Once a valid ticker is identified, the app predicts and returns the stock's closing price for the following trading day.
How we built it
We built the app using Python with Flask for the backend, and connected it to a frontend built with good old base HTML, CSS, and JavaScript.
Challenges we ran into
One issue we faced was not knowing any flask, we had to learn flask while implementing for our app. Another issue was not understanding how LSTMs work and in further research we realised it was easier to implement a simpler model (i.e. an RNN) and we also didn't understand how ARIMA worked so we had to learn that on the spot however the stats model library meant we didn't really need to understand it anyway (and to be honest I still don't really understand how it works) as it is an easy blackbox to implement to check our RNN.
Accomplishments that we're proud of
Without meaning to sound too corny we are proud of the entire project as it allowed us to learn a lot of new skills and kept me up at night learning how to use flask and implement an RNN.
What we learned
-Web development skills -Improved knowledge of web dev and JS/HTML/CSS -How RNNs work
What's next for Stock Price Predictor
-We are going to connect it to a database where it can store users -We might implement a history system -And potentially use more models to make a better prediction -Add functionality to allow the user to change the dates graphed and to choose other predictions (i.e. predict Low in 2 days as currently you can only get the Close the next day)
Log in or sign up for Devpost to join the conversation.