From pages to patterns: Towards extracting catalytic knowledge from structure and text for transition-metal complexes and metal-organic frameworks

• Examine differences in catalysts studied by different communities, leading to fragmented structural databases, distinct formats, and unique challenges for finite versus periodic systems. • Revisit how simple analyses over large sets of data may unveil correlative trends or reveal anthropogenically biased assumptions made on smaller data sets. • Emphasize the importance of generating and curating experimental structural and property data in a FAIR (findable, accessible, interoperable, and reusable) format. • Demonstrate how linking structural data with properties appropriately is insightful over large sets. • Highlight how increasingly sophisticated generative models on text and structural modalities will help future catalysis studies.