Fuzzy numbers - How many Twitter users are truly in the Middle East?

Yesterday, SpotOn Public Relations released a widely-cited report that detailed the amount of Twitter users in the Middle East and North Africa. We published two articles based on the figures; one by media reporter Keach Hagey, "Middle East Twitter users sees rapid growth", and by technology reporter Tom Gara, "Facts show Twitter yet to lure Middle East users".

In short, the numbers gave a brief introspective view at how the micro-blogging service has grown in the area, while ensuring that if there's any local companies out there looking to do some "internet marketing", it looks like SpotOn knows their audience.

But I was skeptical. I spent over five years of my undergraduate education analysing some pretty intense statistics and I just didn't buy SpotOn's methodology, only cited in their press release as: " based on the agency's survey, tracking and analysis of registered Twitter accounts across the Middle East and Africa region."

To confirm these numbers, I turned to Sysomos, a Canadian-based start-up, who hold a patent-pending technology born from a university research study on analysing the social web in real-time. The company made its mark among the Twitterati after publishing a fairly comprehensive global study on the usage of Twitter, full of fancy graphics and all that.

It apparently took the brains over at Sysomos quite a bit of time to process the data. I received the figures from Sysomos late last night, past our paper's deadline, from company spokesman and Twitter expert in his own right, Mark Evans.

The results comparing both firms' figures are after the jump:

What can we see from above is that while the numbers are close, they aren't perfectly spot on (pardon the pun). Although some areas such as Tunisia represent a 1 per cent difference, there is a 54 per cent discrepancy with Saudi Arabia's numbers. Closer to home, Sysomos found more than 700 more Twitter users in the UAE than SpotOn's numbers. Even the final total has a 10 per cent difference of about 1,400 users. Given the overall population of the data, that's quite the tolerance, if you're a stats geek.

What explains the discrepancy? I'll let SpotOn comment below on how they tracked their data. Meanwhile, Sysomos taps into Twitter's API (the data "firehose" itself) to yield is results and are funded by the venture capital arm of the Ontario provincial government. I think they're legit.

In any case, the numbers reinforce my blog colleague's view that Twitter just isn't that important here in the Middle East. As he puts it: