Non-fungible tokens, or NFTs, have been in the focus of emerging digital asset transactions, have gained a lot of attention of art investors, and are among the fastest growing Web 3.0 and crypto products. The popularity of NFTs has started with a generative art collection—CryptoPunks from Larva Lab in 2017, and the total market capitalization has been over US$ 41 billion since 2021. We present a short history sketch of this NFT art phenomenon and investigate the properties of CryptoPunks via their flattened pixel vector. This leads to very high-dimensional data, which we represent via modern VizTech, in order to detect clusters and similarities among the digital assets. Employing the UMAP technique, we are able to attach to and patterns and elements. The Uniform Manifold Approximation and Projection (UMAP) is explained in detail and compared with other localizing techniques like t-SNE. The two-dimensional projections of UMAP disclose clusters among these CryptoPunks which one can associate with prices and traits.