NFT data ingestion is challenging due to the lack of centralized or verified data, as well as the lack of uniform standards for basic definitions. Therefore, the data analysis we see today is often distorted or wrong.

To fully understand this concept let’s examine the four levels of NFT data (a basic understanding of smart contracts is needed — https://www.youtube.com/watch?v=ZE2HxTmxfrI)

1. Blockchain nodes

Blockchain nodes are the moderators who create the infrastructure of a decentralized network. Their primary function is to maintain the consensus of a blockchain’s public ledger. It is the basic level at which data can be extracted on NFTs. The node can retrieve data of a smart contract including transactions detail, balance, and basic information. Each transaction between two parties includes — timestamp, transaction action, buyer wallet, seller wallet, tokens transferred, transaction value, and fees.

Main difficulties:

1. The data structure can change from each blockchain and currently there are more than 20 NFT-support layers 1/layer 2 blockchain solutions (However 80% of the trading volume is on the Ethereum network).

2. Most of the deployed smart contracts are not verified and can be deployed differently. It creates a huge database that is not filtered and contains unnecessary information.

3. There is more than one standard for NFTs (ERC1155\ERC721).

Examples for NFT infrastructure API:

Blockdaemon API

Alchemy API

Etherscan API

2. Marketplaces

Types of applications (dApps) that can operate autonomously, typically through the use of unique smart contracts. An NFT marketplace enables users to list, offer, sell and buy NFTs. Each marketplace collects data on the trading activity for the NFT project. Therefore, we can extract some interesting trading activities like floor prices, royalties fee, social media links, best collection offer unique owners, listing ratio, last sales etc.

Main difficulties:

1. There are dozens of marketplaces, with each having different data on each project (However, ~80% of ETH trading is on OpenSea)

2. There are “shared contracts”. One contract containing multiple projects and artists. Thus, the data points cannot be separated for each project.

3. There is no regulation on wash trading and spam projects.

4. Many main marketplaces are luck of APIs like SuperRare Labs

5. OTCs trades are not regulated

Examples for NFT marketplaces API:

OpenSea API

Looks Rare API

Magic Eden API

3. On-Chain Analytical Tools

These tools allow tracking, analyzing, and exploring NFT projects and often provide premium API. This level is where most of the data history calculations are done like 30D, 7D, and 1D floor price, and the “Value Proposition” API is also added. That is, a data point that distinguishes a data provider from the rest like NFT valuation, Blue-chip holders, whale tracking, rarity, liquidity and more.

Main difficulties:

1. Each API is expensive
2. There are no standards for each data point. For example, by what formula are the rarities of NFTs estimated?
3. The raw data they rely on is not reliable enough

Examples for NFT analytical API:

NFTGo API

Nansen API

Nefertiti API

4. Off-Chain Social data

NFTs are social assets and are highly affected by the project’s community and team. Like in the classical capital markets, NFT prices are sensitive to the news of the moment with most of the social activity happening on social media platforms like discord and Twitter. The social KPI is probably determined by the number of followers, influencers, related posts, unique profile pictures, and community reach.

Main difficulties:

1. Bots, Pump-And-Dump, and disinformation
2. Many announcements are irrelevant and it is very difficult to track the one that does.
3. Social data is off-chain data, therefore doesn’t have public records.

Examples for NFT Social API:

LunarCrush API

NFTInspect (Future API)

Future Thoughts

Misunderstanding of the 4 levels of NFT data can lead to misinterpretation and from there to bad decision-making. As I have shown in this paper, the possibility to build models and KPIs on top of the current NFT databases is almost nonexistent. Without proper standards and collaborations, blockchain decentralization, transparency and accessibility actually can cause investors to be misled.

I believe that this problem can be mitigated with 4 complementary approaches:

1. Verification— I think NFT marketplaces and wallets can verify projects against spam, wash trading and rug pulls projects. I don’t think the idea of gatekeepers goes against the blockchain ideology. As there are ERC standards that guide the development of smart contracts, it is possible to establish a verification standard that every NFT project must comply with. One way to deal with centralization is for those gatekeepers to serve as governance DAOs.
2. Redundancy— Each analytical tool tries to provide the most powerful API and ignores the others. However, to create a better ecosystem we need to allow the implementation of open source protocols. A good example is the OpenRarity protocol designed to be a “single source of truth for NFT rarity”. The way to choose the “winning” protocol is through platforms like Smoothie that allow voting. Another insight is to try to sell individual endpoints to enable the aggregation of data from a variety of tools.
3. Reconciliation — The beauty of an all-encompassing data aggregation protocol is reconciling on-chain data with off-chain data: companies will be able to customize the data links to make it work. This includes data cleansing. and cross-chain/marketplace aggregation.
4. Institutionalization— As in the traditional financial markets, institutions from the fields of education, research and finance can greatly contribute to the understanding and development of the data levels of the NFT space. Institutions have money, knowledge and qualitative human capital capable of building a broad and agreed-upon infrastructure and driving adoption on a wider scale.