Abstract
Non-Fungible Tokens (NFTs) are a developing area in the market of digital assets. NFTs represent digital or real-world items like artwork, gaming collectibles and real estate. We aim to study the daily working of NFT market and its interaction with cryptocurrency (Ether and Bitcoin) and search interest.Our approach involves identification of models encompassing both global and local feature importance. Various regression methods are utilized to determine the feature importance and select the predictive features effectively. Moreover, this study explores the relationship between search interest and weekly NFT sales and vice versa, to comprehend how public interest impacts the NFT market. Lastly, anomalies in daily sales are detected and analysed using STL Decomposition and SHAPely.The study reveals that intrinsic sales attributes and trade profits drive daily NFT sales, with positive sentiment significantly impacting Ethereum volatility and NFT sales. External factors like NFT supply, Ether price, and trade profits also influence anomalies. Positive sentiment significantly shapes crypto and NFT market dynamics.