Inspiration
Our inspiration came from a simple thought. Crypto and betting tend to go hand in hand together; so why is there no app that can do that safely and without being sketchy?
How does it work?
To start when you first log into the app you are redirected to the log-in page. Where you have the option to log in or create a new account. Now once you make an account on PredicTurf you will be prompted to log in to your Near wallet. After which you can finally enter the dashboard. There are a few key things you can do here. The first thing you might see is the match predictions. These predictions are made by a custom machine-learning model trained on the newest data available. This lets it make the most accurate estimates. Towards the right of the match predictions is the leaderboard; which contains information on who has made the highest bet. This is used to engage the audience. Bellow both of these are your betting amount options. There are six quick presets and then a custom amount option available. Once an amount and team have been chosen the user can Successfully click the pay icon which will make the server interact with their Near wallet deducting the amount. Furthermore, the server will give a notification to the user if they lost or won their bet. The user also will receive stickers based on the amount they bet. These stickers are made using a GPT-4 and would be accessible in the user profile and also through their Adobe account. Essentially making an NFT wallet with Adobe Express add-ons.
Challenges we ran into
We ran into many challenges. To start even getting an idea was a challenge. But the biggest obstacles we came to overcome were implementing the use of Near's wallet, Uploading images to Adobe Express, and training our machine learning model.
Things we Learned
Even before we started, we knew this was going to be a challenge as half of us were high schoolers and the entirety of our specialty was in web dev. But we persisted and learned to use Near's API, Neurelo cloud API, Web 3, How to build and train a machine learning model, and Adobe Express add-ons API.
Things we would improve