-
-
BTC query
-
SOL query
Inspiration
We wanted to make a simple machine learning web app to predict whether or not the price of a crypto coin is going to go up or down after the next 24 hours.
What it does
The machine learning algorithm we used is a Random Forest Classifier trained on historical data gathered from yahoo finance. The web-app allows you to enter the stock ticker for any crypto coin and get a prediction and confidence value based on the machine learning algorithm
How we built it
Small python files to create a FastAPI to query our ML algorithm, with a lightweight and simple front-end to allow simple inputs for coin tickers.
Challenges we ran into
We initially built our mvp on our own without much structure. This led to a poorly structured project that was almost impossible to connect the front-end and back-end together. Around 1 AM we pivoted to prompting chat-gpt to build a skeleton framework of our mvp, where we built it up from there. We didn't have much experience with either web-dev or ML so this was certainly frustrating at times, but we learned a lot about both.
Accomplishments that we're proud of
We were able to get a simple mvp finished, even after facing some difficult challenges
What we learned
We learned a lot about machine learning algorithms, like how they are built, trained, and tweaked for optimal performance given certain metrics.
What's next for Crypto Currency Predictor
Rebuilding it from the ground up after establishing a more solid foundation, and planning a lot more thoroughly than we did during the Hackathon.
Log in or sign up for Devpost to join the conversation.