FOSS4G NA 2024

KnowSTAC: Framework for Creating Dynamic and Knowledge-Interoperable STAC Indexes
09-10, 11:30–12:00 (America/Chicago), Grand C

KnowSTAC is a framework for creating dynamic STAC indexes that organize data not according to how they are published as collections or catalogs, but according to on-demand queries that can crosscut domains and organizations for improved collaboration.


KnowSTAC is a framework for creating dynamic SpatioTemporal Asset (STAC) indexes that organize data not according to how they are published as collections or catalogs, but according to on-demand, user-defined and ad-hoc queries that can crosscut domains and organizations. Collections can be dynamically created and shared according to semantic classification, location (including hierarchies and spatial knowledge graphs), and time so that they can be discoverable for collaboration. For example, the results of a dynamic search including geological classification, vegetation, platform, sensor type, organizational hierarchy, geographic hierarchy, and/or temporal validity collectively might represent an answer to a question that, by making discoverable, could be incorporated into other collections to answer new questions.

KnowSTAC is being developed for GeoPlatform.gov to support the national spatial data infrastructure (NSDI) mandate and to explore transitions to a geospatial knowledge infrastructure (GKI).

SDIs are supplier-driven data networks, traditional data analytics, and services that are single direction and created for a general market. They are not sophisticated enough to allow for processing random, unpredictable, and context-dependent queries across multiple domains. A GKI is a knowledge-sharing paradigm that is client-driven by on-demand questions of individuals and the relationships between those questions across domains. Rather than publishing data, a move beyond SDI to GKI enables answers to be published that are interoperable with other questions.

Our GKI approach involves using spatial knowledge graphs and ontology registries to facilitate semantic interoperability. Ultimately, our aim is to enable interoperability across different technologies. For example, the results of a STAC index query should be available in RDF so that their relationship with other spatial knowledge graphs can be discoverable using GeoSPARQL or used to bring geospatial awareness to Large Language Models.