OSNews

Their numbers jumped because they came out with a new phone, and people wanted the new one. They do this once a year. in order to get any sense of reality, you would need to start from the launch of the iPhone 4 and go until just before the release of the 4S. Since new Android phones are released about every half hour, their release cycle is not an issue. If you want to see how the 4S affected things, you would need to look at a period from now back to the same amount of time after the 4 was released (to account for the shiny new gadget syndrome.)

None of this accounts for how long an Android user keeps the same phone as compared to how long and Apple user keeps the same phone, so even getting realistic sales numbers does not equate to market share.

If you want market share, you need to look at devices currently in use. In the the USA this is possible, since the carriers generally know what phone you are using, but in the land of unlocked phones, even that does not work.

If there were a website that were heavily trafficked, and neither hip (Apple heavy) nor nerdy (Android heavy) then maybe that would work. Maybe the server logs for Wikipedia or Facebook are the best real indicator.