Skip to content.

TADA Contents

Find topic


Search text

Web tools

Help

Tools

       Analysis Tool Bar  +
Main > Projects > TAPoRware > TappingWebServices

Turning Web Services into TAPoRTools

This is to document some ideas about turning certain web services into useful TAPoRTools.

Possible Candidates

OpenCalais

The Calais web service automatically attaches rich semantic metadata to the content you submit – in well under a second. Using natural language processing, machine learning and other methods, Calais categorizes and links your document with entities (people, places, organizations, etc.), facts (person ‘x’ works for company ‘y’), and events (person ‘z’ was appointed chairman of company ‘y’ on date ‘x’). The metadata results are stored centrally and returned to you as industry-standard RDF constructs accompanied by a Globally Unique Identifier (GUID). Using the Calais GUID, any downstream consumer is able to retrieve this metadata via a simple call to Calais.
Developer Docs are available here.

  • I have been talking to Abhay Kumar about his Ruby-based web app that demonstrates OpenCalais very aptly.
  • A PHP Class is available from Dan Grossman.
  • A great example of how one can carry out faceted browsing of semantically tagged news articles is available at LinkedFacts.com.
  • A good way to see the bare and immediate results of a call to Calais is via this form. p.s. I really like their little activity thermometer/submit button. It probably breaks a bunch of UI rules, but I find it rather cool nonetheless.

One of the ideas that Shawn has been noodling on is to be able to have OC provide a tagset for a doc, potentially under JiTR, and be able to the determine like documents based on their tagsets. Depending on the extent of the tags applied, greater semantic distinction could be metatized for a document, and used to differentiate amongst a large collection, but also identify salient similarities that might otherwise have gone unnoticed. I always seem to refer to this as digital DNA.

If we are able to identify semantic similarities, then we can offer a tool much like Amazon's preference engine (which largely just relies on maintaining a repository of transactions and breadcrumb - i.e. matching user interaction). OpenCalais could take this a step beyond. This isn't to say that taking a word frequency as a fingerprint and finding similar word frequency distributions amongst oher documents might not in itself lend towards this.

ManyEyes

Many Eyes is a social networking site built around sharing datasets and building visualisations of that data using a set of tools provided in-site. Tools currently allow for working with textual and numeric data and present the visualisations using a mixture of Flash and Java. One of the strength of the site is their explanation of the appropriateness of types of visualisations to data.
Many Eyes can only barely be described as a web service. Although specific datasets and visualisations are tokenized for sharing, you cannot make a call to a particular visualisation generation tool and have it return a visualisation. We might be able to beg our Monk friends to share with us an indication that Many Eyes might be opening up their service.
Their sole nod to the demand to be able to work with contributions has been to embed a static link jpeg on a user's page.

-- ShawnDay - 24 Apr 2008


Use this box to quickly add a comment to the page.

more options...