NFT Collectors Database

Alex Kosenkov
Art Keepers
Published in
2 min readAug 27, 2022


We’re glad to announce the release of NFT collectors preference database to our community members.

For each collector we include a wallet address, twitter handle (if available) and preference estimate towards NFT collections based on their purchase history.

In order to create it we’ve sifted through all Ethereum and Tezos NFT transactions, scraped twitter handles from public sources and trained a collaborative filtering model to estimate the collector preferences.

The resulting database is available (after request) in Google Spreadsheet format and comes in a few flavours:

Preferences towards Art Blocks collections

Preferences towards fxhash collections

Preferences towards all popular NFT collections (for Art Blocks collectors)

If you’re savvy in python, here’s a simple example showing which collections are most correlated to Fidenza.

>>> import pandas as pd
>>> df = pd.read_csv('collectors_database.csv')
>>> df.columns
Index(['address', 'twitter', 'similar_collectors',
... ])
>>> df_ = df.drop(columns=['address','twitter','similar_buyers'])
CREATOR_0xa7d8d9ef8d8ce8992df33d8b8cf4aebabd5bd270:100 0.620428
CREATOR_0xa7d8d9ef8d8ce8992df33d8b8cf4aebabd5bd270:13 0.627430
CREATOR_0xa7d8d9ef8d8ce8992df33d8b8cf4aebabd5bd270:130 0.627943
CREATOR_0xa7d8d9ef8d8ce8992df33d8b8cf4aebabd5bd270:23 0.677429
CREATOR_0xa7d8d9ef8d8ce8992df33d8b8cf4aebabd5bd270:89 0.700285
CREATOR_0xa7d8d9ef8d8ce8992df33d8b8cf4aebabd5bd270:78 1.000000
Name: CREATOR_0xa7d8d9ef8d8ce8992df33d8b8cf4aebabd5bd270:78, dtype: float64
Usage example: the collections correlated to Fidenza

We hope this data will help you to increase efficiency of your social campaigns and find even more beautiful artworks!

Don’t hesitate to reach out to us directly if you have any question via And make sure to subscribe for our curated digests!