FAIR for Experimentalists: Let's be Visual - Catalysis Lab Data meet Knowledge Graphs

Experimental catalysis research produces complex, heterogeneous data that is often difficult to make FAIR (Findable, Accessible, Interoperable, and Reusable), especially from the per-spective of experimentalists. Our community has long debated what FAIR truly means in our daily lab work [1]. That changed when we focused on having our workflows modelled visually. This contribution presents how we transformed the experimental procedures from our recently published study on catalytic oxidative cleavage of vicinal diols [2] into a structured, visual, and FAIR representation. This example served as a proof-of-concept for modeling laboratory pro-cesses in a reusable, semantically rich way. By breaking down the experimental workflow into typed entities (e.g., processes, samples, and equipment), we created a flexible data model that communicates scientific knowledge more clearly than traditional text-based methods. The true test of a resource's "FAIR-ness" is whether its successful reuse by another FAIR resource. That's why we based our modeling on existing controlled vocabularies like Voc4Cat [3], and terms from ChEBI, RXNO, CHMO, [4] as well as ORCiD, and ROR. Where neces-sary, we assigned new persistent identifiers (PIDs) to key components and samples, laying the groundwork for interoperability and reuse. The importance of PIDs is recognized by NFDI by the recently formed PID Coordination Hub, PID4NFDI. Our modeling process begins with asking researchers to describe their workflows in as few words as possible, resulting in simple, intuitive mind maps (Figure 1). In modeling the proce-dures from our research paper, we identified 31 processes, 35 sample states (including trans-formations), and 11 pieces of equipment. The resulting model was initially stored as a labeled property graph (1000+ nodes and edges) and is currently being converted to RDF to align with FAIR principles for communication. The granularity of information is adjustable via built-in hier-archies, allowing different levels of detail depending on audience or application. The effectiveness of this approach was validated internally: lab members quickly grasped the modeled workflows, and even non-experts could understand the methodology more easily than through the original publication text. This increases transparency, accelerates onboard-ing, improves reproducibility, and opens the door for future machine-readable applications such as AI-driven hypothesis generation or automated lab reporting. This work highlights how FAIR practices can be integrated into everyday experimental re-search without requiring a background in data science. Our approach is not about forcing standardization from the outside, but about empowering experimentalists to put their knowledge at the center, by structuring it with the help of RDM tools. In our view, reusability is the ultimate test of FAIR-ness. By demonstrating how our own vocabulary terms have been reused across multiple workflows, and including external semantic resources, we advocate for this as a sustainable and scalable model. In catalysis, as an intrinsically interdisciplinary field, visual, interoperable workflows could ena-ble more efficient collaboration across disciplines. Visual FAIR helps us not just to publish da-ta, but to grow smarter, more interconnected research ecosystems.