Daniel Garijo (University of Southern California), Maximiliano Osorio (University of Southern California), Deborah Khider (University of Southern California), Varun Ratnakar (University of Southern California), and Yolanda Gil (University of Southern California)
Software is crucial for understanding, reusing and reproducing scientific results. Software is often stored in code repositories, which may contain human readable instructions necessary to use it and set it up. However, a significant amount of time is usually required to understand how to invoke a software, prepare data for its execution and to reuse it in combination with other software. In this paper we introduce OKG-Soft, an open knowledge graph that describes software in a machine readable manner and a framework to annotate, query, explore and curate software metadata. OKG-Soft emphasizes the ability to compose software, proposing an ontology to describe inputs and outputs of software components and their expected variables. We demonstrate the usefulness of OKG-Soft with two applications in the environmental and social sciences: a tool for exploring software models and a portal that exploits the contents of the graph to combine climate, hydrology, agriculture and economic models.