Advancements in synthesis capabilities often create the problem of an inability to test all the newly available materials, thus motivating methods relating synthesis conditions, structure, and properties. A crucial ingredient to these methods is a quantitative representation of the materials in question, which is especially challenging for materials of increasing complexity. In many disciplines, complex systems are represented using graph theory (GT). However, despite its unique capabilities, the use of GT to complex materials is relatively new. In this presentation, we present our open-source package, StructuralGT, for the development of structure-property relationships for complex materials. We show how the API was developed to combine the speed of a C backend with the usability of a Python frontend. We show how StructuralGT has been applied to develop diverse structure-property relationships for three experimental systems: electrical properties of multilayer conductive films, mechanical properties of strut lattices, and chiroptical properties of complex chiral nanodendrimers. Finally, we give heuristics for computational researchers looking to develop GT structure-property relationships for their own complex materials, including use of matrix methods, shortest path finding, and insights from finite-difference time domain simulations.
