12 Jan

#protestbaku hashtag analysis

Facebook Twitter Email Pinterest Reddit Tumblr

EDIT 5pm Pacific, 5am Baku – Since this has been so popular, I’ve made a more inclusive version here.

Thankfully the hashtag #protestbaku caught on and I was able to archive the tweets for analysis. I helped create it about 3 hours before the protest started.


So here’s a map of the hashtag (thanks Marc Smith!)

nodexl map

link to full

What does this tangle of wires mean? These are clusters of people who are communicating with one another. You can see that there are some pretty tight networks here. (And if you’re familiar with the Azerbaijani Twitter scene, the clusters shouldn’t surprise you too much.)

Group G1 on the far left consists of a lot of Azerbaijanis who don’t live in Azerbaijan anymore. A lot of the URLs that they were tweeting can be considered ones that allow for them to “keep up” with what happened during the protest.

Group G2 are people that, more or less, were on the ground and centers around Adnan Hajizada. Their URLs tweeted were more of photos than liveblogs or videos. They also were more likely to talk about logistics – where police were, as well as break news.

Group G3 includes people that I’m not really familiar with. They linked to a lot of YouTube videos though and used words that are affiliated with the pro-government groups.

Group G4 is very much pro-government people. Their hashtags used were against the protest, more or less. Rauf Mardiyev, the leader of the pro-government youth group is the center of that.

Group G5 is more opposition but seems to center around Baxtiyar Hajiyev, another opposition leader. I’m not really sure why Baxtiyar’s group is so distance from Adnan’s. Perhaps Adnan (and Emin Milli) have a different Twitter following because of their relatively early fame and frequency of tweeting in English. Looking at Baxtiyar’s twitter confirms that he mostly tweets in Azerbaijani.

Groups G6 and G7 seems to be unimportant.

You can download an Excel spreadsheet of all of this and play around with things like word frequency — no statistical knowledge required.

I also did a TweetReach analysis of #protestbaku. You can do the last 50 tweets for free. I bought a set of 1500 tweets though. Here’s what that found.

– There were 188,045 Twitter users who were exposed to a tweet that contained the #protestbaku hashtag.
– There were over a million “impressions” – 1,075,736 to be exact. That means, essentially, someone’s eyes potentially read #protestbaku over a million times in the last day.

The twitter who got the most exposure was Jamal Ali with 50,300 impressions. The most retweeted and most mentioned was @muntezir.

With that I have a list of tweets, most popular, etc. if anyone is interested.

So what? What does this mean?

A few things…

1. In this protest, Twitter “mattered” – but it mattered in different ways for different people. For people on the ground, it was used a bit for logistics, but mostly for getting information out. For people not present, it was a way to spread information.

2. There is polarization. This isn’t surprising. I guess the amount of guff from the pro-government twitter crowd was a little new for me. They’ve gotten louder in the past year, but their simultaneous tweeting during the protest and use of the hashtag was interesting, to say the least.

3. Personally I “met” a lot of new Azerbaijani tweeters today through this hashtag. Were these people whom I just don’t run into otherwise? Were they people that don’t tweet usually but did for this event? It goes to show that a hashtag has the potential to bring people together.

4. Speaking to the clusters of groups, I wonder if someone who would be interested in pulling people together for greater collection action could focus on those people who are between networks?

I’d love to hear other thoughts on this. I usually keep my comments off to avoid spam, but I’ll turn it on now!