Spatial Data Search: a View from the Space Station

Presentation | Presented

  • Jo Walsh, OSGeo

This talk emerges from a research project with Terradue on best practise for metadata reuse. The development of "ontologies", data classification schemes, is controversial; both ad-hoc and formal approaches significantly fail. Keyword and text search is over-emphasised in the literature on "discovery services". Why bother to describe data more than necessary?

A future direction in spatial data search is outlined, in which

* search is based more on data contribution as a starting point
* search and contribution run in a GIS client, not a web browser
* "model-exchange based query" can look at structural similarity, scale, proximity to a sample data set
* past usage patterns can help to filter future queries
* participation is tied to the act of data distribution and annotation

Lots of simple strategies for making data easier to find, when plugged together, yield quite powerful results. This is the "space station view" of spatial data search. To start with is a simple metadata backbone, and a few easy pieces.

For more information: