Developers‎ > ‎

Thomson Reuters Open Calais API

posted Nov 7, 2013, 4:43 AM by Mpelembe Admin   [ updated Nov 7, 2013, 5:21 AM ]

 Open Calais is a free Web Service from Thomson Reuters that helps you automatically 
analyze your unstructured textual documents – by converting the unstructured text into a 
meaningful structured text - metadata. We call it –aboutness. You can send document by 
document using the HTTP Restful API and receive back the metadata for every document 
you send. The metadata is a list of entities, events, relations, facts and topics. You can learn 
more by reading here
 
To get more familiar with Open Calais and how to submit documents please read 
 
To start using Open Calais please read this Get Started page. 
 
There are several code samples to use in Java, python, .Net, Ruby and more here
 
Few tips on some of the cool capabilities of Open Calais: 
Open Calais provides for every entity extracted – the relevance score – how relevant this 
entity to the document. This is a very important capability to help you filter the value from 
the huge data out there. For example many documents mention Google but much less of the 
many documents are really about Google. Relevance score helps you find the documents 
that are about particular entity. 
SocialTags are categories applied to a document that are coming from the Wikipedia 
taxonomy. Very interesting. Each category is relevant to the document (although the term 
might not be explicitly mentioned in the document) and these categories are also articles in 
Wikipedia. 
Generic Relations helps you find any kind of relationships between entities to entities or to 
events. 
Disambiguating entities: Companies, Geographies, people – are disambiguated and resolved 
to a normalized name, increasing the accuracy of the extraction and providing a consisted 
tagging across many different documents. 
 
Open Calais is one of the first Linked Data services out there. You can read more about 
Linked Data in Open Calais here. It is linked to Freebase, DBpedia, IMDB and more. 
 
For questions – please email us
 
 
  
 32
Useful Stuff and Link
Comments