Mining the tweets

25 November 2010 | By David Wilson

Although it’s widely touted as a rich source of instantly updated information that can keep its readers abreast of an incredibly wide variety of topics, it’s taken me a while to fully understand the relevance of the Twitter website.

And it’s easy to see why. After all, it’s not intuitively obvious what practical purpose there is to posting a hundred or so words up onto the internet in a so-called ‘tweet’. Unless, of course, you are asking to be fined £1,000 under section 127 of the 2003 UK Communications Act or sentenced to a one-year sojourn in a Chinese penal complex.

But the other day I came to the realisation that the true worth of this rather peculiar website does not simply lie with the information that individuals are posting up on their tweets on a daily basis, but rather in the way that manufacturers might exploit the information found there to their own advantage.

That’s right. With so many people using Twitter, it’s inevitable that, among all the comments regarding the death of Dobby, the marriage of HRH Prince William and Miss Catherine Middleton, and the sonorous rendition of The Beatles’ tune ‘Get back’ by Wagner on the X Factor, there’s also a plethora of information that relates specifically to people’s opinion of manufacturers and the products they produce.

And, if such data could be extracted from the noise of the other less industry-specific issues, the information certainly be used to many manufacturing companies’ advantage.

It’s not difficult to see how this might be achieved. Today, many companies are already using data-mining software techniques to extract patterns from data that can then be used to pinpoint and help rectify problems that might arise in their manufacturing and service businesses.

So there’s good reason to believe that those companies presently using such software would be only too delighted to be able to analyse data mined from Twitter feeds as well. Such an analysis would provide them with an additional stream of information that could also help them to identify customer problems and settle manufacturing issues.

The wonderful thing about the data on the Twitter website is that it is free to view − and free to mine as well. As such, companies no longer need to employ expensive research teams to gain some insight into the acceptance of their products in the market. Instead, all they would need to do is purchase or develop a sophisticated enough data-mining package that could extract the important information they desire.

Interestingly enough, when such data-mining processes become more widely adopted by manufacturers − as they surely will be  − the folks that contribute to such websites with their opinions of companies and their product failings will become no more than unpaid researchers for the companies themselves. Of course, that’s a fact that none of the social websites are likely to be boasting to their subscribers about!

Dave Wilson
Editor, Engineeringtalk

Dave’s comments form part of the weekly Engineeringtalk newsletter, which also includes a round-up of the latest engineering products and services. To subscribe click here