The problem Nft Similarity Detective solves

An investigator can use this for telling whether NFTs in a collection, or amongts multiple collections, are really unique or not. They can display different NFT's easily, and then get to have machine learning do the rest for them. However, they would still need to be informed about the domain, for the machine learning outcomes weren't automatically interpreted on our end. Instead, we present them to the user for them to derive meaning out of it.

Maybe NFT platforms could use a more developed version of this. They could have all, or most of the 'relevant' NFT's embeddings present on a database, and compare and contrast when a new NFT is identified to the system. However, our aimed customer for this project were people who'd like to find something and tweet about it instead :D

Challenges we ran into

We ran into some API limits with ethscan and arbiscan API's, however we solved this on later stages by downloading the ABI ourselves and manually putting which we thought would be sufficient for the scope of this hackathon.