The British Geological Survey (BGS) recently organized an internal hackathon to investigate ‘semantic pathways’ that would make BGS resources more accessible. BGS’ Rachel Heaven led ‘team semantic search’ that set out to implement a ‘semantically and spatially intelligent search service’ and to ‘cut through the tangle of complex geoscience terminology and improve navigation between information resources.’
BGS’ textual information includes observations and interpretations that accompany the hardcopy 2D geological map. Current digital technology has failings in searching across such data sets and may be ‘divorced’ from the documentary evidence they are based on because of a lack of ‘provenance’ information.
Vanilla search simply matches a sequence of letters. Semantic search adds domain-specific terminology and an ‘ontology’ of relationships between terms and concepts. BGS here leverages its collaboration with the W3C’s Data Activity initiative. In fact BGS has been working to codify geoscience terminology since before the semantic web, notably via the IUGS-CGI geoscience terminology working Group.
The hackathon team used ElasticSearch to index its textual material and then used PL/SQL to create a gazetteer and query expansion tool. The BGS’ chrono-stratigraphic ontology was thrown into the blender along scripts to retrieve terms from web pages. The team demonstrated text-matched results from the indexed document collection and also manufactured a virtual cross section from disparate BGS resources including the Vale of York 3D model. Read Heaven’s blog here.
© Oil IT Journal - all rights reserved.