Wednesday, 18 March 2015

Language Spy

    At the developer conference I wrote about in a previous post, I had the usual name badge on a lanyard round my neck. Under my name: "Language Spy".

    A significant number of the people I talked to assumed it was my job title. Pity, that, as it's the name of my startup.

    Some time in early February I moved on from the place I called work for four and a half years, gave notice on my flat, and set to coding. As a dotcommer I never move away from the startup mentality, so to me it seems like the most natural thing in the world to be doing.
    I'm a geek-of-all-tech-trades, but I work in the language business. I came to it via a very convoluted path through electronics, computer games, the web, and then search engines. What I'm doing now builds on one of my fascinations, hobbies, and side projects of many years standing, the statistical analysis of large bodies of text. I gather huge amounts of language, tell my computer to make observations on it I couldn't possibly do myself, and then write code to render the result in a human-readable form.
    I've got several products in the pipeline, but my first is live on the web and involves time-series political data. If you want to know which of the party leaders is making waves in the UK election campaign, or which of the 2016 US presidential hopefuls is most affected by whatever banana skins Fox News toss at them, it's the tool for you. I am waiting for the general election with interest, because it'll allow me to test my theory that I can use corpus data to call the result.

The URL?

    I kinda wish I had adopted it as a job title now.


  1. Nice URL. Look forward to some face to face language in very near future.

  2. Nice URL. Look forward to some face to face language in very near future.

  3. I do hope all goes well, this election will be a difficult one to call so if you do that will certainly say something about your system.

  4. fascinating, as a political junkie I look forward to reading about your results

  5. On news this morning it sounded as if they had been using something like this to find most popular words in use...