We’ve been busy as bees adding some nifty new features to our NexaMaster platform lately.
Not an actual picture of our developers. We don’t let them go outside.
By working with Macromeasures, we’ve successfully integrated demographics and other goodies into our system. In all Twitter and Gnip projects regardless of language, you will find the following reports concerning the Twitter account holders in your project:
- Primary language
- Platform (eg, iPhone, Android, Computer, etc.)
- Age group
- TV Shows of interest
- General interests
- Location: cities and countries
How can we know all this you say? Well, it is thanks to a combination of explicit and implicit mentions: what is in their public bio, what they tweet about, what their handle is, what networks they belong to, what language they are using in their tweets, and how they act. For example, if you tweet an awful lot about going to East Side Mario’s it’s a fair bet you’re in the greater Toronto area. Elementary, my dear Watson.
Pipe is optional but recommended.
Of course, it is hard to assess all the handles for everything – ergo, the reports count only those handles for which stats are pretty certain. For example, the gender report is between 95 and 97% accurate on the handles it scores.
We’ve built our own state of the art sentiment engine. Sentiment is now available in all free Twitter and Gnip projects, in English only for now. It’s a pretty simple classification (negative, positive, neutral).
Ask your friendly neighbourhood Guardian Angel (me) for more information if you have questions about the reports.