Abnormal growth in the share of Internet Explorer 7 in the post-Soviet space. Someone can explain?

Question

Browsers

Today I was leafing through browser statistics according to Statcounter and noticed a terrible anomaly with the share of IE7, which manifests itself in many post-Soviet countries, but is not seen anywhere else.
It looks something like this:

There was a natural predictable decline until October 2012, but then a rise suddenly began, which can hardly be called a statistical error or a local anomaly, because it continues to this day, already for half a year.
It became interesting for me to look at the bigger picture and it turns out that the global, pan-European, pan-Asian trends continue to show a smooth predictable decline. We also looked at trends for various countries of the ball, taken absolutely at random: China, India, Sweden, Slovakia, South Africa, Colombia, Bangladesh, USA. There is a beautiful downward trend everywhere.

But if you walk through the countries of the former USSR, then almost everywhere you notice the same anomalous rise:
Ukraine
Russia
Moldova
Kazakhstan
Tajikistan
Latvia
Lithuania
Belarus

I have only one question: what does this mean?

Answer 1

Maybe bots? Some kind of conspiracy? Or a joke? No matter how strange it may sound, but I have not yet found more adequate options.
Due to the intermediateness of the 7th version (XP -> IE6, Win7 -> IE8), it is very doubtful that this is a live activity. It is unlikely that users would massively download the 7th version.

Answer 2

Bots?

Answer 3

Really curious picture.
Although liveinternet shows a natural decline, except for the last week.

Answer 4

Statcounter is not reliable and terribly buggy statistics system.
marketsshare.hitslink.com/?source=NANL
At one time, of course, when it was necessary to prove that the share of IE6 was already miserable, I used its statistics. But even then I understood that he was lying godlessly. Yes, and there were studies (somewhere there was an article on Habré) that the collection methodology from Statcounter is not correct.