Molecules in the interstellar medium act as powerful probes for physical and chemical properties in astrophysical environments. To this end, building up a known inventory of molecules “accumulates tools in a chest”, and when combined with chemical models provides a comprehensive picture of complex astrophysical processes. To date, however, there is no systematic way to construct and grow a network of molecules both in an observational and a chemical network sense; we are unable to readily identify which molecules are missing, and which will provide critical constraints on chemical models. Our approach, called the Molecular Recommender System (MRS), combines big data chemistry with astrochemistry, taking known molecules found in an astrophysical environment and predicts similar species that might be abundant using Gaussian Processes. To demonstrate our methodology, we apply it to study the well-known dark molecular cloud, the Taurus Molecular cloud (TMC-1) where the inventory is substantial and well-characterized. The methods involve translating molecules to a machine representation that will be able to be compared. The unsupervised system learns from observed column densities and suggests species from three large molecule databases from cheminformatics and astrochemistry, including the Kinetic Database for Astrochemistry (KIDA), based on the magnitude and uncertainty of their modeled column densities. The system is able produce target molecules that can be considered in chemical models for TMC-1 and in future observations, and also provides candidates for unidentified lines of spectra. The SAO REU program is funded in part by the National Science Foundation REU and Department of Defense ASSURE programs under NSF Grant no. AST-1852268, and by the Smithsonian Institution.