Spyglass combines data analytics on NFTs with a UI focused on the collector's needs to provide an all-in-one place to quickly view and find the best NFTs. Buying an NFT can be chaotic, traders need to reference various tools to arrive at a valuation and have hundreds of tabs open. It's difficult to stay organized! Spyglass uses machine learning to generate predictive pricing, and also shows whether an NFT is over/under valued by trait. This analytics data paired with a snappy UI that allows users to use keyboard shortcuts to quickly curate from a collection, makes Spyglass an important tool for any NFT trader.

Spyglass  showcase

How it's made

We used Zeplin to design the project and built the front-end with React and Redux. Node, Express, and MongoDB were used for the REST API server which syncs asset and collection data from the OpenSea API and appends it with price predictions and trait valuation scores from a machine learning model. The machine learning model was built with Python, Jupyter Notebook, and scikit-learn. We trained the model on the full sample of a collection, performing a 80:20 split. The algorithm used was the RandomForestRegressor.