Tag: microblogging

State of the Calgary Twittersphere – August 2009

Welcome to the sixth State of the Calgary Twittersphere, my look at the intersection of Twitter and Calgary. You can see last month’s stats here.

The source of the data this month is the same as last, primarily Twitter Search but also from Twitter directly. If a user has his or her location set to Calgary, Airdrie, Okotoks, Cochrane, Strathmore, or matching lat/long coordinates, they are considered a Calgarian. If a tweet is “about Calgary” it contains either the word Calgary, the #yyc hashtag, or both.

For August 2009:

# of local users: 6852 (an increase of 226 from July)

To clarify, that means there were 6852 users who posted at least one tweet in August 2009 with their location set to something that makes them an Calgarian as described above. This number should be treated as a minimum – there are probably many more Calgary users without their location set or that were not captured for some other reason.

Here are the rest of the stats for August 2009:

# of tweets by local users: 244330
# of tweets by local users containing #yyc: 5463 (2.2%)
# of tweets by local users that were replies: 75678 (31.0%)
# of tweets by local users containing links: 67295 (27.5%)
# of tweets by local users that were retweets: 9466 (3.9%)
# of tweets by local users that were twooshes: 6945 (2.8%)

Here are the numbers above in graphic form:

Here are the top clients used by local users for posting updates. UberTwitter doesn’t seem to be as popular in Calgary as it is in Edmonton:

Some other interesting stats for the month:

Final Thoughts

Like the Edmonton numbers, growth in August for Calgary was minimal. It’s the summer, after all! I suspect that growth in September will be higher.

Since March, I have identified just over 13,500 local users. The number above, 6852, is how many of them were active in August.

State of the Edmonton Twittersphere – August 2009

Welcome to the eighth State of the Edmonton Twittersphere, my look at the intersection of Twitter and Edmonton. You can see last month’s stats here.

The source of the data this month is the same as last, primarily Twitter Search but also from Twitter directly. If a user has his or her location set to Edmonton, St. Albert, Sherwood Park, Leduc, Nisku, Stony Plain, Fort Saskatchewan, Beaumont, Spruce Grove, or matching lat/long coordinates, they are considered an Edmontonian.

For August 2009:

# of local users: 5365 (an increase of 232 from July)

To clarify, that means there were 5365 users who posted at least one tweet in August 2009 with their location set to something that makes them an Edmontonian as described above. This number should be treated as a minimum – there are probably many more Edmonton users without their location set.

Here are the rest of the stats for August 2009:

# of tweets by local users: 206318
# of tweets by local users containing #yeg: 12808 (6.2%)
# of tweets by local users that were replies: 72060 (34.9%)
# of tweets by local users containing links: 41015 (19.9%)
# of tweets by local users that were retweets: 9947 (4.8%)
# of tweets by local users that were twooshes: 6828 (3.3%)

Here are the numbers above in graphic form:

Here are the top clients used by local users for posting updates. I’m pretty impressed by the rise of UberTwitter:

And by popular demand, a new stat this month! Here are the number of local users by community (besides Edmonton itself) for August:

Keep in mind that is based entirely on the text inside the location field of the user’s profile, and is definitely a minimum. Still, gives you an idea of the breakdown.

Some other interesting stats for the month:

 

Final Thoughts

A little bit of growth this month, but not much. It’s the summer, after all! Hopefully that means everyone was outside enjoying everything Edmonton has to offer. Growth in September should be higher.

Since January, I have identified just over 10,400 local users. The number above, 5365, is how many of them were active in August.

I’ll have the Calgary stats posted in the next couple of days!

State of the Calgary Twittersphere – July 2009

Welcome to the fifth State of the Calgary Twittersphere, my look at the intersection of Twitter and Calgary. You can see last month’s stats here.

Apologies for the delay in getting this one posted. The source of the data this month has changed, slightly. In addition to Twitter Search, I added more data directly from Twitter itself. A total of 16,082 tweets were pulled from user timelines directly because they did not exist in Twitter Search.

User identification remains the same. If a user has his or her location set to Calgary, Airdrie, Okotoks, Cochrane, Strathmore, or matching lat/long coordinates, they are considered a Calgarian. If a tweet is “about Calgary” it contains either the word Calgary, the #yyc hashtag, or both.

For July 2009:

# of local users: 6626 (an increase of 980 from June)

To clarify, that means there were 6626 users who posted at least one tweet in July 2009 with their location set to something that makes them an Calgarian as described above. This number should be treated as a minimum – there are probably many more Calgary users without their location set or that were not captured for some other reason.

I’m happy with my previous stats posts existing as point-in-time snapshots. That said, I am continually improving my data set, making it more accurate over time. I am still happy with the numbers I posted for March through May – they are reasonably accurate (within about 100 users). The number I posted for June however, 5646 users, is not accurate. It was actually 6506, which means this month actually increased by 120 users (however I left the 980 above for consistency).

Here are the rest of the stats for July 2009:

# of tweets by local users: 238525
# of tweets by local users containing #yyc: 5028 (2.1%)
# of tweets by local users that were replies: 72969 (30.6%)
# of tweets by local users containing links: 58297 (24.4%)
# of tweets by local users that were retweets: 7713 (3.2%)
# of tweets by local users that were twooshes: 8159 (3.4%)

Here are the numbers above in graphic form:

Here are the top clients used by local users for posting updates. You’ll notice that API is now identified separately from web:

Some other interesting stats for the month:

Analysis

There wasn’t much growth in July, which isn’t altogether surprising considering it’s the summer. I suspect things will pick up again in the fall. I’m still fascinated by the difference in tagged tweets between Calgary and Edmonton (you can see Edmonton’s stats for July here).

Since I started recording the stats for Calgary back in March, I have identified just over 12,500 local users. The number above, 6626, is how many of them were active in July.

Twestival Local is happening in Calgary on September 12th! Click here for more info.

State of the Edmonton Twittersphere – July 2009

Welcome to the seventh State of the Edmonton Twittersphere, my look at the intersection of Twitter and Edmonton. You can see last month’s stats here.

The source of the data this month has changed, slightly. In addition to Twitter Search, I added more data directly from Twitter itself. A total of 19,199 tweets were pulled from user timelines directly because they did not exist in Twitter Search.

User identification remains the same. If a user has his or her location set to Edmonton, St. Albert, Sherwood Park, Leduc, Nisku, Stony Plain, Fort Saskatchewan, Beaumont, Spruce Grove, or matching lat/long coordinates, they are considered an Edmontonian.

For July 2009:

# of local users: 5133 (an increase of 1491 from June)

To clarify, that means there were 5133 users who posted at least one tweet in July 2009 with their location set to something that makes them an Edmontonian as described above. This number should be treated as a minimum – there are probably many more Edmonton users without their location set.

I’m happy with my previous stats posts existing as point-in-time snapshots. That said, I am continually improving my data set, making it more accurate over time. I am still happy with the numbers I posted for January through May – they are reasonably accurate (within about 100 users). The number I posted for June however, 3642 users, is not accurate. It was actually 4581, which means this month actually increased by 552 users (however I left the 1491 above for consistency).

Here’s a graph to illustrate local user growth in 2009:

Here are the rest of the stats for July 2009:

# of tweets by local users: 210162
# of tweets by local users containing #yeg: 12879 (6.1%)
# of tweets by local users that were replies: 73119 (34.8%)
# of tweets by local users containing links: 43702 (20.8%)
# of tweets by local users that were retweets: 8393 (4.0%)
# of tweets by local users that were twooshes: 7185 (3.4%)

Once again, the number of tweets containing #yeg actually includes all local hashtags, such as #yegtransit or #yegfood (though most often they are used together anyway).

Here are the numbers above in graphic form (not corrected for June):

Here are the top clients used by local users for posting updates. You’ll notice that API is now identified separately from web:

Some other interesting stats for the month:

Finally, here are the top ten users in Edmonton by followers (most followers first): Pat_Lorna, revtrev, DrBarryLycka, LesM, subunit1, garrymullen, redneckmommy, mastermaq, NHL_Oilers, dan_cote.

Analysis

Looks like we’re getting some growth again, which is a little surprising considering it’s the summer. A couple of things helped to boost July’s stats, including the City Centre Airport issue and the big summer storms. I counted 1464 tweets posted by local users with the #ecca hashtag in July, and I already posted some stats on the storm here. Additionally, July 8th (which is when the Council meeting on the airport was held) was the second busiest day of the month.

Since January, I have identified just over 9200 local users. The number above, 5133, is how many of them were active in July.

I’ll post the Calgary stats as soon as I can – still some manual processing to do, so it might be next week before I get it all finished.

Twitter promoting Search despite major issues

Yesterday Twitter launched a new home page that puts more emphasis on search and trending topics. There’s a nice big search box on top, with up-to-date, daily, and weekly trends underneath. The aesthetic is different from the rest of the site however (you don’t see any of this if you’re logged in), so don’t be surprised to see additional changes in the coming weeks.

If you enter a query or click on a trending topic, the search results appear below. It looks a lot like Twitter Search. Some of the improvements include a description of what the trending topics are (Hell’s Kitchen was given the description “A reality television cooking competition”) and search tips appear in a little box on the right.

I don’t think the new design should be a surprise to anyone – it has been clear for quite some time that Twitter Search is important.

What’s surprising is that they’re promoting search even though it has major issues:

  • Stale Results: Twitter itself has become very stable lately, but the same cannot be said for Twitter Search. Results routinely become stale, sometimes for as long as an hour or two (so the newest tweets to show up in the results were posted an hour or two ago). For a real-time search engine, the stale results issue happens surprisingly frequently.
  • Missing Tweets: Over the last few weeks I’ve noticed that the number of missing tweets has increased (though I think it has always been an intermittent problem). It used to be that I could enter my username and see all replies at Twitter Search, but lately I can’t. Some tweets simply don’t appear in the Twitter Search index. I’ve submitted a support request about this, but have not heard anything back yet.
  • Other Intermittent Issues: There are a few good reasons that someone might not appear in search results (such as if they have a private account) but lately Twitter has had issues keeping the index up-to-date with new accounts.
  • Lack of Innovation: With the exception of adding the “source” property to search results, Twitter has done very little to improve the service they purchased a little over a year ago. Real-time search is new and ripe for innovation, but Twitter doesn’t seem interested. One of the oldest quirks is that user IDs returned from Twitter Search don’t match up with user IDs at Twitter itself. This is scheduled to be fixed in the next version of the API, but it’s not clear when that will happen.

Worst of all, Twitter has been terrible at communicating about the above issues. The Twitter Status blog is never updated when search results go stale, and very little has been shared regarding the future direction of Twitter Search.

The good news is that Twitter is finally starting to acknowledge that they need to improve search. Last night, Biz wrote: “We have a lot of work to do when it comes to the quality of our search results and trend analysis…”

Search is vitally important to Twitter, and I want to see them succeed. If they don’t address the above issues however, someone else is going to come along and steal their thunder.

The Last 24 Hours on Twitter: Storm in Edmonton

We’ve been dealing with a pretty massive storm here in Edmonton for the last 24 hours or so (Edmonton Journal, iNews880). At one point last night, every fire hall in the city was dealing with some kind of storm-related issue. Epcor has been taxed dealing with power outages, and I’ve heard nothing but sirens downtown.

As expected, many Edmontonians turned to Twitter to communicate about the storm. At one point last night, #yeg was the 6th most talked about item on all of Twitter. I was aggregating tweets together on a livestream that you can look at here.

Here’s what the last 24 hours looks like for local Twitter users:

Since noon yesterday, there have been just under 8000 tweets posted by local users, most of them with the #yeg hashtag. Other popular tags include #yegweather and #yegstorm.

Storm in EdmontonStorm in Edmonton

There have been dozens of photos and videos posted online too. The two above are a couple that I took last night before things got really crazy. The two most popular TwitPic’d photos from last night were:

Lots of cleanup happening around the city today, though the wind and rain are still quite intense at times. I’ll have more on this tomorrow!

State of the Calgary Twittersphere – June 2009

Welcome to the fourth State of the Calgary Twittersphere, my look at the intersection of Twitter and Calgary. You can see last month’s stats here.

The source of the data this month remains the same – Twitter Search. If a user has his or her location set to Calgary, Airdrie, Okotoks, Cochrane, Strathmore, or matching lat/long coordinates, they are considered a Calgarian. If a tweet is “about Calgary” it contains either the word Calgary, the #yyc hashtag, or both.

For June 2009:

# of local users: 5646 (a decrease of 215 from May)

To clarify, that means there were 5646 users who posted at least one tweet in June 2009 with their location set to something that makes them a Calgarian as described above. This number should be treated as a minimum – there are probably many more Calgary users without their location set or that were not captured for some other reason. See additional analysis below.

# of tweets by local users: 207245
# of tweets by local users containing #yyc: 4269 (2.1%)
# of tweets by local users that were replies: 68737 (33.2%)
# of tweets by local users containing links: 48114 (23.2%)
# of tweets by local users that were retweets: 6637 (3.2%)
# of tweets by local users that were twooshes: 8865 (4.3%)

Here are the numbers above in graphic form:

Here are the top clients used by local users for posting updates (remember that web includes all unidentified API calls too). A notable new entry is UberTwitter:

Since I’ve been playing with some temperature data from Environment Canada lately, I decided to plot the average temperature against the number of tweets for each day in June:

Do people tweet more when it’s colder outside? I’m not sure the graph really tells us. What’s clear however is that people tweet less on the weekends!

Some other interesting stats for the month:

Finally, here are the top ten users in Calgary by followers (most followers first): douglasi, OksanaIrwin, ahhhgolf, twitty7x, strategicsense, CrazyMechanic, broatch, nolanmatthias, ThankASoldier, tessaru

Analysis

Like the Edmonton stats for this month, the Calgary stats don’t look that impressive. Some people have abandoned Twitter, it’s the summer, and as was pointed out in the comments on the Edmonton post, the Iran Election probably impacted my stats as some users changed their location to Tehran.

There are some encouraging signs, however. The increase in tagged tweets this month surprised me – maybe #yyc will try to catch up to #yeg! Also, since I started recording the stats for Calgary back in March, I have identified just over 11,000 local users. The number above, 5646, is how many of them were active in June.

I didn’t monitor things as closely in June as I do normally, so there’s a chance I might not have captured all of the data. In particular, I’m not sure if June 13th/14th was really that low or if there was a data import error of some kind. Either way, I think the data above represents the overall trends accurately.

I’ll be making some changes to hopefully boost the accuracy of the numbers in July.

State of the Edmonton Twittersphere – June 2009

Welcome to the sixth State of the Edmonton Twittersphere, my look at the intersection of Twitter and Edmonton. You can see last month’s stats here.

The source of the data this month remains the same – Twitter Search. If a user has his or her location set to Edmonton, St. Albert, Sherwood Park, Leduc, Nisku, Stony Plain, Fort Saskatchewan, Beaumont, Spruce Grove, or matching lat/long coordinates, they are considered an Edmontonian. If a tweet is “about Edmonton” it contains either the word Edmonton, the #yeg hashtag, or both.

For June 2009:

# of local users: 3642 (a decrease of 871 from May)

To clarify, that means there were 3642 users who posted at least one tweet in June 2009 with their location set to something that makes them an Edmontonian as described above. This number should be treated as a minimum – there are probably many more Edmonton users without their location set. See additional analysis below.

# of tweets by local users: 172295
# of tweets by local users containing #yeg: 9411 (5.5%)
# of tweets by local users that were replies: 64281 (37.3%)
# of tweets by local users containing links: 31937 (18.5%)
# of tweets by local users that were retweets: 7532 (4.4%)
# of tweets by local users that were twooshes: 6453 (3.7%)

Once again, the number of tweets containing #yeg actually includes all local hashtags, such as #yegtransit or #yegfood (though most often they are used together anyway).

Here are the numbers above in graphic form:

Here are the top clients used by local users for posting updates (remember that web includes all unidentified API calls too, though they are changing that). A notable new entry this month is UberTwitter:

I’ve been playing with some temperature data from Environment Canada lately, and it occurred to me that I could visualize it against my Twitter data. Here is the number of tweets posted and the average temperature for each day in June:

Do people post more when it’s warmer or when it’s colder? The only thing I can be sure of is that there are always less tweets posted on the weekend. Beyond that, there seemed to be more tweets posted at the beginning of the month when temperatures were cooler, but obviously this is in no way scientific!

Some other interesting stats for the month:

Finally, here are the top ten users in Edmonton by followers (most followers first): Pat_Lorna, revtrev, LesM, subunit1, garrymullen, mastermaq, redneckmommy, NHL_Oilers, karlkovacs, dan_cote. It would take a lot more time to generate, but a top ten list ranked by local followers would probably look much different.

Analysis

This month’s stats don’t look very good! The numbers haven’t been this low since March. There are definitely users who have abandoned Twitter, but I think the time of year probably has an effect as well. Presumably people are outside enjoying the weather, tweeting less often! A preliminary look at the numbers for July suggests to expect more of the same for next month’s stats.

The good news is that since January, I have identified nearly 9000 local users. The number above, 3642, is how many of them were active in June.

I didn’t monitor things as closely in June as I do normally, so there’s a chance I might not have captured all of the data. In particular, I’m not sure if June 14th was really that low or if there was a data import error of some kind. Either way, I think the data above represents the overall trends accurately.

I should have the stats for Calgary up by Thursday. Stay tuned!

UPDATE: The Calgary stats for June are here.

#yeg – Year in Review

It’s hard to believe that just a single year has gone by since the first tweet with the #yeg hashtag. It has become such a core part of the Edmonton Twitter community that it’s difficult to imagine what microblogging would be like without it!

If I remember correctly, I found about the #yyc hashtag while I was in Calgary for BarCampCalgary2 on June 14th, 2008. I learned from @wintr that a few Calgarians had started using the hashtag to tweet about things related to their city. I thought it might be a good idea to do something similar here in Edmonton. This is what I wrote:

#Edmonton Twitterers…our neighbours to the south have started using the hashtag #yyc to self organize…cool idea we should borrow? #yeg

The rest, as they say, is history.

A few highlights from the past year:

  • #yeg was mentioned on CTV when Carrie Doll did a story on Twitter at the beginning of March.
  • #yeg trended to #5 on Twitter Search on March 24th as Edmontonians were tweeting about the fire on Rice Howard Way.
  • No less than ten related hashtags have been actively used, such as #yegtransit and #yegfood.

I recall at some point a discussion about how to pronounce #yeg. I think most say “y-egg” though some say “y-e-g”.

Unfortunately, I don’t have a lot of stats for the hashtag before January 2009. My best guess is that 1488 tweets were posted containing the #yeg hashtag from November 20th to December 31st. For this calendar year however, I have good numbers. The stats below differ from my monthly stats posts in two ways: they are global, so include tweets posted by non-locals, and they do not contain related tags such as #yegfood.

A total of 32053 tweets have been posted in 2009 (as of noon today) containing the #yeg hashtag. Here’s that number broken down by month:

On average, 188 tweets are posted each day containing the #yeg hashtag.

One thing I’ve always wondered is where the hashtag is most often placed inside each tweet. Here’s the breakdown:

I think it’s great that the #yeg hashtag has become so established in the Edmonton Twitter community, and I look forward to its continued evolution. Thanks everyone for making #yeg so great!

State of the Calgary Twittersphere – May 2009

Welcome to the third State of the Calgary Twittersphere, my look at the intersection of Twitter and Calgary. You can see last month’s stats here. Apologies for posting this so far into June, but my travelling over the last few weeks made it difficult to do. Better late than never!

The source of the data this month remains the same – Twitter Search. If a user has his or her location set to Calgary, Airdrie, Okotoks, Cochrane, Strathmore, or matching lat/long coordinates, they are considered a Calgarian. If a tweet is “about Calgary” it contains either the word Calgary, the #yyc hashtag, or both.

I’m not sure what happened on May 15th, but I seem to be missing data for that day (the same system imports data for both Calgary and Edmonton, and the Edmonton stats were unaffected).

For May 2009:

# of local users: 5861 (an increase of 210 over April)

To clarify, that means there were 5861 users who posted at least one tweet in May 2009 with their location set to something that makes them a Calgarian as described above. This number should be treated as a minimum – there are probably many more Calgary users without their location set or that were not captured for some other reason.

# of tweets by local users: 209260
# of tweets by local users containing #yyc: 3228 (1.5%)
# of tweets by local users that were replies: 74979 (35.8%)
# of tweets by local users containing links: 43354 (20.7%)
# of tweets by local users that were retweets: 6593 (3.2%)
# of tweets by local users that were twooshes: 9459 (4.5%)

Here are the numbers above in graphic form:

Here are the number of local users created per day in May, using the best available data from Twitter (it seems to be really unreliable, I will probably drop this stat next month):

Here are the top clients used by local users for posting updates (remember that web includes all unidentified API calls too):

Some other interesting stats for the month:

  • The ten most active local users (most tweets first): mrrocknroll, burstingenergy, C_DIG, strategicsense, birdalert, wikkiwild1, bish0p, aprilcandy70, SalBarguil, iKasperr
  • Just over 52% of all local tweets were posted between 9 AM and 5 PM.
  • Local users posted roughly 4.7 tweets per minute in May, compared with 4.6 per minute in April.
  • The day with the most local tweets posted was May 5th at 9114. On average, 6824 local tweets were posted each day (compared to 6561 in April).
  • Of the 74979 replies posted by local users this month, 16327 or 22% were to other local users.
  • A total of 847 users posted 50 times or more in May. In comparison, 1085 users posted just once.

Here are the ten most replied to local users for May: C_DIG, mrrocknroll, ubershmoo, caninez, amacisaac, jonincalgary, aNorthernSoul, eviltera, darylcognito, CdnCowgirl

Analysis

Like the Edmonton stats for May, it might appear at first as though Twitter growth in Calgary has stagnated. Instead, I think there’s just some attrition taking place. There were 1879 users who posted at least one tweet in May that did not post a tweet in April (so presumably they are new). Likewise, there were 1635 users who posted at least one tweet in April that did not post a tweet in May (so presumably they abandoned the service).

In total, I have identified just over 8000 local users since March – the number above, 5861, is how many of them are currently active.