Inspiration
Born from a need to ball on a budget.
What it does
Uses aggregated price data from StockX product sales to learn from price trends and make future market predictions up to one month in advance.
How I built it
A Python machine learning script collects historical price data from StockX.com using their RESTFUL API. From there, it calculates an intelligent price prediction one month in the future using a _ scikit-learn _ linear support vector regression model trained on the collected data from the platform. The price data is then passed into Firebase for storage, which then communicates with our Android app for display to the user in chart format.
Challenges I ran into
Selecting a machine learning model which best generalized to the various StockX products and their diverse price patterns by far proved to be one of the greatest challenges and consumed the majority of our time.
Accomplishments that I'm proud of
Actually building the project from idea to full working prototype given the time provided, testing through the night, the teamwork and perseverance required to realize our idea.
What I learned
How to implement an effective machine learning model and apply it given our problem domain, and how to feed that ML data to a mobile app.
What's next for FootForward
Expansion and upgrades to our Firebase and its connection to the StockX API, allowing us to search for any item listing on the site and then running the ML script on a remote server or cloud service.
Log in or sign up for Devpost to join the conversation.