Linking Open Government Data

A well-known benefit of using Linked Data principles when publishing government data is the ease with which the resulting data may be integrated and linked with diverse other datasets. Realizing such benefits ultimately hinges on unambiguously identifying common entities within these datasets. From a practical standpoint this means identifying which literals represent instances of entities --- US states, US government agencies, toxic chemicals, crops, etc --- and associating canonical URIs with those occurrences across datasets. To facilitate interlinking, it is then helpful to relate these literals and our newly-minted canonical identifiers to other identifiers in common usage --- DBPedia data pages, information pages used in various industries, etc --- to further facilaite interlinking with other datasets.

The TWC LOGD Instance Hub Project seeks to demonstrate the utility of this approach for entities occurring across diverse government datasets. First, we have identified several categories of entities from Data.gov datasets. We have designed canonical URIs based on a draft set of principles that may be applied across a wide range of government data sources (agencies, etc). And we have provided links to synonymous identifiers for these entities in common use in other systems to facilitate interlinking.

TWC has established a draft set of URI Design Principles that government agencies may choose to adopt as they establish authoritative, canonical URIs for entities of special concern within their datasets. The TWC LOGD Instance Hub demonstrates how to associate these canonical URIs with instances of entities that occur as literals or as URIs from other naming systems.

The following table provides links to instance data for a number of categories derived from US Government data via Data.gov. Click on a category to browse these instances or download the RDF data (in Turtle).