There is an exponential explosion in the amount of materials data being generated due to the advancements of measurement equipment, high-throughput testing protocols, simula- tions, and data generation initiatives. However, the existence of more data is not directly correlated to a better understanding of material systems, as it generally does not meet the findable, accessible, interoperable, and reusable (FAIR) principles. The heterogeneous na- ture of the data, disparate sources, and inconsistent use of schemas (data models) has made the generation of FAIR and navigable materials datasets elusive. Therefore, analyzing mul- tiple sources of data and extracting insights from multiple projects requires a great deal of conceptual background and preprocessing, which slows research or prevents analysis alto- gether. This work develops the materials informatics knowledge graph to integrate a diverse and heterogeneous set of datasets. It incorporates a flexible ontological schema which captures the important relationships needed to represent challenging and complex materials concepts such as process-structure-property linkages. The schema is validated by demonstrating the ability to incorporate a hierarchy of material, process, and data classes needed to represent the diverse datasets, mining the datasets using a knowledge representation framework, and gaining insights about the datasets using only an understanding of the unifying schema and graph learning tools. ; Ph.D.