Catalyst

For this, hackathon we have created price feed that can uses hedonic linear regressions to determine the value of the CryptoPunk market as a whole and can be used to price the value of each individual Punk in real-time based on on-chain data.

NFTs suffer from a notable liquidity issue which makes evaluating the value of an NFT difficult to determine and in turn difficult to sell. Since NFTs as an asset are most similar to real estate we decided to apply a similar pricing model that is traditionally used for real estate derivatives and valuations.

Hedonic regression is used in hedonic pricing models and is commonly applied in real estate, retail, and economics. Hedonic pricing is a revealed-preference method used in economics and consumer science to determine the relative importance of the variables which affect the price of or demand for a good or service. For example, if the price of a house is determined by different characteristics, like the number of bedrooms, the number of bathrooms, proximity to schools, etc., regression analysis can be used to determine the relative importance of each variable.

Real estate and NFTs are similar assets due to their poor liquidity, the value of CryptoPunks can be derived by determining the value of each attribute. For Punks, we analyzed historical transactions, the crypto marketcap at the time of sale, and the attributes such as Full-Body type and # of attributes. By applying linear regressions; we can estimate the weighted value of each analyzed variable to determine the value of the entire Punk market and each individual Punk.

This model can be applied to the entire NFT market and set a baseline pricing for the NFT market as a whole. In addition, the valuation can be useful for implementing collateralized NFT loans. The value of the entire Punk market is useful for the creation of a synthetic derivative that can bring liquidity, simplify investments and introduce a short-selling mechanism.

Our most notable accomplishment was solving the issue of $0 transactions for cryptopunks when a bid is accepted. This was done by analyzing the time of the transaction for a Punk and correlating that to an accepted offer which allowed us to determine the price of the transaction. Many CryptoPunk subgraphs exist and we couldn't find one that resolved this issue for the developers.

Next we had to determine how best to take into consideration the attributes and body type of a CryptoPunk to apply our hedonic linear regression model. Although Punks have over 30 different attributes we determined that the Full Body Type(ex. Female-Dark), attribute count and the value of the cryptomarket at time of sale were what really determined the price of a Punk.

In addition we have another github for the frontend https://github.com/lyst-finance/hackmoney-frontend

I'm very impressed by our pricing model for individual Punks, this can easily be applied to other markets and can be a useful calculator that can establish a baseline pricing for NFT valuations. These valuations could then be useful for other financial tools such as NFT lending where the collateral value of the NFT is established with real-time data and fluctuates with the value of the NFT market.

We had to use NodeJs, React, and the Graph to query Punk transactions and feed them to our index pricing.