Skip to content

Automate metadata with opencalais

sebbacon edited this page Jan 12, 2011 · 1 revision

In order to facilitate cross-linking between requests, authorities and so on, we could consider automated metadata extraction using OpenCalais.

The drawback is that this only works in English, French, and Spanish.

As an example, the following entities were extracted from a request about business rates (chosen at random). I particularly like the tags, and the fact we can extract company names and localities from the data.

##Tags

  • Oxfordshire 1
  • United Kingdom 1
  • Geography of England 1
  • Property taxes 2
  • Local taxation 2
  • Oxford 2
  • Taxation in the United Kingdom 2
  • Abingdon Road 2
  • Business rates in England and Wales 2
  • Littlemore 2

Entities

  • Currency: GBP (0.32)
  • Person: Chris Brewer (0.32)
  • Organization: Oxford City Council (0.80)
  • Company: RO Developments Ltd (0.10)
  • Person: Michael Newman (0.38)
  • Facility: Town Hall (0.34)
  • City: Oxford (0.76)
  • Position: Information Officer (0.36)
  • Organization: Corporate Secretariat (0.38)
  • Position: Manager (0.38)
  • Facility: YAMANOUCHI RESEARCH INSTITUTE (0.29)
  • Company: Yamanouchi UK Ltd (0.60)
Clone this wiki locally