This is the third part in my State of the Edmonton Twittersphere for 2011 – you can read the overview post here. As we saw in that entry, local users posted more than 2.5 million tweets containing links. They also posted more than 3 million tweets containing hashtags. Let’s take a closer look at what exactly we tweeted about in 2011.
Here’s a breakdown of the number of tweets posted per month containing links. As expected, it trends up:
More interesting than the number of links is what those links are for! Of course, most links are hidden behind shortened URLs using a service like TinyURL. That makes it difficult to see the final destination of the link. Additionally, Twitter rolled out its own URL shortener in August which automatically wrapped all links with t.co. So to get a sense of what the links look like, I examined the data from January through August and excluded any t.co wrapped links. That gives us this:
As you can see, bit.ly is by far the most popular URL shortening service. Looking at the top ten, we can see that Tumblr, Facebook, Foursquare, and YouTube account for a large number of the links we post on Twitter every day. So, we use Twitter to link to other social networking sites!
Here are the top ten domains (excluding t.co):
Other popular domains included instagr.am, goo.gl, and dlvr.it.
Perhaps more useful than links for determining what we tweeted about are hashtags. After all, the hashtag was created as a way to “categorize” tweets. In 2011, local users posted more than 3 million tweets containing hashtags. The most commonly used hashtag, by a longshot, was #yeg. Here is a word cloud of the most commonly used hashtags:
Local users used nearly 600,000 different hashtags in 2011. The shortest was just one character, excluding the hash, and the longest was 139 characters. The average hashtag was 14 characters long. There were seven hashtags that filled the full 140 characters of a tweet (when you include the #) and all but one of them had to do with using a long hashtag (the outlier was an extended hey!).
This should give you a sense of the diversity of hashtags used, and also of the prevalence of #yeg:
Removing #yeg from the word cloud allows us to get a better sense of the other top tags:
Here are the top ten hashtags excluding #yeg:
The only one that surprises me there is #customer. I have no idea why that hashtag was so popular! Given that it hasn’t been used recently, I suspect a bot may have helped boost its use at some point throughout the year.
I have always been fascinated by the number of local “subtags” as I call them, or hashtags that start with #yeg. In 2011, local users posted more than 3600 different hashtags that started with #yeg. The average length, excluding the hash, was 11 characters. The longest was 52 characters. Appropriately, it was about the weather – #yegohmygoodnesscantbelievewinterisherehowwillimanage.
Here’s a look at the top #yeg-related hashtags:
Here are the top ten #yeg-related hashtags:
So what did we talk about on Twitter in 2011? The same stuff we talk about everywhere else – food, politics, music, traffic, etc. The list above is pretty similar to the list from 2010, with #yegfood on top and a number of political tags like #yegcc, #yegarena, and #yegfed.
Here’s another way to visualize how popular those ten hashtags were, when compared with all other #yeg-related tags:
And here’s a look at the next most commonly used tags after excluding the top ten:
I imagine #yegdt, #yegmedia, and #yegbiz will all climb the charts in 2012!
2011 in Review
This year I have broken the report into sections:
This entry brings the report to a close. I hope you found it useful. Thanks for reading!
Whew! Thanks for your exhaustive work on giving us a look at the #yeg Twittersphere.
Thanks for reading!
Hello friend!!
I enjoyed reading your article and it gives all the information on a helpful topic. Actually, I was not at all aware of this. Thanks for sharing it with us.
Incredible, Mack!
Love the way you think – and the window you provide for me into this world of thought. Vely vely intelesting.
🙂
Valerie
Yes fantastic work collecting this data Mack!
Although I’m curious if you experienced any anomalies that might be associated with inconstancies in twitter’s API with respect to geocodes? I’ve been noticing issues with several clients getting inconsistent geocode search results leading me to believe the prob may be twitter. Even search.twitter.com was acting weird!