Software:Zanran

From HandWiki

Zanran is a search engine for data and statistics.[1][2] Zanran's focus is on finding graphs, charts and tables on the Internet, which distinguishes it from other search engines such as Google, Bing, etc.[3][4] Unlike a typical search engine, the results—graphs, tables, etc.—can be previewed by mouse-hovering over the thumbnails.[3][5]

History

In 2006, the founders, Dr Yves Dassas and Jon Goldhill, started developing the technology that makes Zanran possible.[2] A limited beta ran, starting in November 2010. The service was launched as a public beta version in May 2011.[3][6][7]

Technology

Zanran has developed two technologies specifically for this application:

  1. Image ‘classification’ is the ability for a computer to decide whether an image is a graph, a pie-chart etc. as opposed to a photograph or a cartoon. The Zanran algorithms work to over 95% accuracy. This is important because most images on the web are not graphs and otherwise there would be a large number of false positives.
  2. Text extraction is the process of taking the most appropriate text to describe the graph. This contrasts with a normal search engine where an entire HTML page might be included.

These processes are the subject of Zanran's UK patent.[8] The image processing in particular takes a great deal of computing power. Zanran runs on the Amazon cloud, and uses hundreds of machines at a time.

The service is English-language only as of December 2011.

The company

Zanran Ltd is based in London, UK. It was financed by the founders prior to a private angel investment round in March 2010.

Other data-search on the Internet

Other specialists in the data-search space include WolframAlpha, Infochimps, and Timetric.

References

External links