October 31, 2007

Long tail of Twitter

The Long Tail is a concept to describe certain business and economic models such as Amazon.com or Netflix. A long tail distribution is characterized by a short head and a heavy tail. In many cases the long tail can make up the majority of the distribution. Businesses such as Amazon.com or Netflix can generate a significant part of their income by selling a greater volume of otherwise hard to find items at small volumes than of popular items at large volumes. The long tail was introduced by Chris Anderson who also wrote a book on this subject : The Long Tail: Why the Future of Business is Selling Less of More.

I wanted to know if the long tail distribution could also be observed in the Twitter world. I started by looking around for reliable quantitative information on Twitter users. There are several sites with top lists of Twitter users. I discovered that these top lists are not 100% reliable (see this blogpost). Furthermore they only cover a top 100 (Twitdir) or a top 150 (Twittown). So I created my own top 500 list.

A long tail distribution follows a powerlaw (or Pareto distribution), which can easily be discovered by looking at a log-log plot of the distribution (both axes on a logarithmic scale) where the actual data points should show up as a straight line in case of a powerlaw. If you click on the graph below you will get a better view at the plots of the distribution of the top 500 of Twitter users by number of following (the number of Twitter users a Twitter user is following), by number of followers (the number of other Twitter users following a Twitter account) and by number of updates (number of Twitter messages published).

The "normal" plots of the distributions show a very short head. The distributions on the log-log plots resemble a straight line, suggesting a long tail.



These graphs are only based on the distribution of a top 500 of Twitter users. There are two problems. In the first place I am not absolutely sure that the lists I compiled are the true lists of top 500 users for the three criteria. It is very likely that I missed several Twitter accounts. Furthermore as there are currently over 500,000 Twitter users, a picture of a top 500 does not tell anything about the rest of the distribution.

Instead of focussing on the complete Twitter community, I zoomed in on two subcommunities, the Twitter community of two countries : Brazil in South America and Belgium in Europe. Looking at the location information in the Twitter profiles I was able to identify 1700 Twitter users in Brazil and 1000 Twitter users in Belgium. European users were among the first to start using Twitter. The popularity of Twitter in South America started later.

Brazil


Belgium


For both countries the observed distributions show similar patterns. A very short head followed by a long tail. The distributions resemble slightly a powerlaw. The correlation coefficients (R-square values) are between 0,76 (Brazil - number of updates) and 0,86 (Brazil - number of followers). The correlation coefficients for the top 500 distributions are even higher : between 0,95 and 0,98 - suggesting a very strong powerlaw relationship.

The long tail concept was introduced to describe how business could sell less to more. As the founders of Twitter haven't yet decided on their business model it is not clear how money can be made on Twitter. How money can made of the long tail of Twitter users remains even a bigger question.

October 25, 2007

TwitterWhere

Twitter is a community of more then 500,000 users. Within this big Twitter community there are many subcommunities. Some Twitters users expressed their need to find Twitter users within their geographical area. But until recently there was no easy way to locate Twitter users from a certain area.

TwitterWhere generates an RSS or XML Feed to filter out Tweets around a certain area. The RSS or XML feeds are based on a city name, state name, or postal code and the range of miles. To access this feed it has to added to an RSS reader (Google Reader, Bloglines, Netvibes, ...). As TwitterWhere is parsing the Twitter public timeline, only Twitter messages from public profiles are known to TwitterWhere. Users who have chosen to keep their message only accessible to their friends (also described private Twitter accounts) will not show up in the RSS of XML feeds, although their profiles often indicate their geographical location.



TwitterWhere is developed by Matt King, an Interactive Developer living in Portland, Oregon. He took only a few hours to get this site running according to a post on his blog. Matt wrote the app in Ruby on Rails.

TwitterWhere was clearly developed with a focus on Twitter users in the United States. Postal codes only work for the USA and the range of the distance is miles, not kilometers.

I subscribed to feeds for several locations in Europe, the continent where I am living. I noticed that the calculation of the distances between two geographical locations can be improved. A few examples to illustrate this.

* Malaga, Spain is within 10 miles of Madrid, Spain according to TwitterWhere, 338.23 miles according to Mapquest
* Bergen-op-zoom, Netherlands is within 50 miles of Brussels, Belgium according to TwitterWhere, but 59.85 miles according to Mapquest
* Lübeck, Germany is within 100 miles of Apeldoorn, Netherlands according to TwitterWhere, but 271.36 miles according to Mapquest
* A Coruña, Spain is within 50 miles of Brussels, Belgium according to TwitterWhere, 1136.53 miles according to Mapquest

TwitterWhere is a nice new tool to support the Twitter community. I hope that the Dual 1.8GHz Powermac G5 server will be capable of handling the traffic as the popularity of this site will surely increase.

October 16, 2007

Twitter Top Lists

If you are looking for different top lists regarding Twitter users you have several possibilities.

The first site with top 100 lists was Twitterholic. Currently there is something wrong with this site, all Twitter users shown on the three lists (top 100 by followers, by friends and by updates) have 1 follower, 1 friend and 1 update.



Looking at the Twitterholic stats for the BBCWorld Twitter account, it looks like Twitterholic is experiencing the problems already for several weeks.



Another site where you can find top 100 lists is Twitdir. The full top 100 lists (followed, updaters, favouriters, followers) are not available as a single page, but you can browse thru these lists in groups of 5 Twitter users.



Top 150 lists can also be found at Twittown (followers, updates and following).



I have noticed that the lists at Twittown and Twitdir do not show the same Twitter accounts. Furthermore the quantitative data shown is not equal for some of the Twitter accounts.

Following


Updates


The current available Twitter top lists from Twitdir and Twittown cannot be considered to be 100% reliable. I can think of only one site that has full access to complete and up to date statistics, namely Twitter Inc themselves. But they do not disclose this kind of information ...

October 12, 2007

Tools to post Twitter messages

Until now I did not encounter an analysis of the most frequently used tools to post Twitter messages. So I launched my own little investigation. I extracted all messages from the homepages of over 2000 randomly chosen public Twitter accounts (there are currently over 500,000 Twitter accounts, as reported in this previous blogpost).

45 different tools were used to post 31,944 Twitter messages on 2,087 Twitter accounts. The 5 most frequently used tools, responsible for 91% of all analyzed Twitter messages, are :
1. web - 60.77%
2. im - 9.87%
3. twitterrific - 9.13%
4. txt - 8.43%
5. twitterfeed - 2.80%

Twitter is multi platform, meaning that messages can be posted thru different channels. Of these channels, web remains the most used channel. The instant messaging interface to Twitter (im) is the second most used channel. Twitterrific is an application that has to be installed on a computer to read and publish twitter messages. On the 4th place we find Twitter messages submitted as text messages from mobiles phones (SMS). Twitterfeed offers the possibility to feed your blog to Twitter, meaning that blog posts will be automatically announced on Twitter.

From each Twitter account maximum 21 Twitter messages were taken into account in this analysis. Surprisingly there were 10 Twitter accounts using up to 5 different channels to publish Twitter messages. Are these Twitter users experimenters ?
66.3% of the Twitter users did use only one tool to publish messages, 34.6% did use two different tools, 12% did use three different tools and finally 2.3% did use 4 different tools.

I suppose that some readers of this blog might be interessed in the full list of tools. The full list is shown in the graph and in the detailed list below.



Tool - Percentage
1. web - 60.77%
2. im - 9.87%
3. twitterrific - 9.13%
4. txt - 8.43%
5. twitterfeed - 2.80%
6. Netvibes - 1.43%
7. TwitterFox - 1.22%
8. Facebook - 1.04%
9. TwitBin - 0.98%
10. Twit - 0.56%
11. Twitter Tools - 0.47%
12. Tweetr - 0.46%
13. foxytunes - 0.42%
14. movatwitter - 0.35%
15. TwitterPod - 0.33%
16. PocketTweets - 0.20%
17. TwitterIrcGateway - 0.15%
18. Hahlo - 0.13%
19. Snitter - 0.13%
20. fring - 0.13%
21. Chattr - 0.10%
22. TwitKu - 0.09%
23. TwitterMail - 0.08%
24. Twitter Line - 0.08%
25. tmitter - 0.08%
26. brabblr - 0.08%
27. iTweet - 0.07%
28. yedda - 0.07%
29. TwitterMSN - 0.05%
30. TwitterGram - 0.03%
31. IMified - 0.03%
32. Twadget - 0.03%
33. Twippera - 0.03%
34. bookey - 0.03%
35. ThinCloud - 0.02%
36. Spaz - 0.02%
37. BlogRovr - 0.02%
38. GtkTwitter - 0.01%
39. twibble - 0.01%
40. Twitterlicious - 0.01%
41. tokotto - 0.01%
42. Twitter Opera widget - 0.01%
43. funP - 0.01%
44. Mobypicture - 0.003%
45. mojungle - 0.003%

October 11, 2007

Twittering without access to Twitter.com

Big companies (and sometimes also smaller companies) have often decided that instant messaging and social networking sites are keeping their employees from the actual work. Hence, they have chosen to block internet access to these instant messaging and social networking sites. As Twitter becomes more popular, Twitter is often also blocked in these companies. What are your options to keep in touch with your Twitter friends during office hours if direct internet access to Twitter.com is blocked ?

There are different Twitter client programs available, but if an employer blocks internet access to specific sites, it is very likely that employees do not have the possiblity to install software on their computers on their own initiative. So I did not take these programs further into account as possible bypasses.

Option 1 : Use your mobile phone to receive and send Twitter messages as text messages (SMS)
Disadvantages

  • If you are following several Twitter users with a high post volume, your mobile phone will continuously receive new messages. Its capacity of storing text messages will soon be reached.
  • Following hyperlinks in text messages is not easy, you have to retype the whole URL in your browser.
  • If the mobile phone you are using is a company phone will your employer accept this usage of the mobile phone ?

Option 2 : Switch to a mobile phone capable of mobile surfing
Disadvantages
  • The size of the screen is rather small.
  • Expensive if you have to pay for such a mobile phone yourself.

Option 3 : Subscribe to the RSS-feed "with friends" from your Twitter account in your favorite web based feedreader such as Bloglines or Google Reader
Disadvantages
  • A web based feedreader does not check every minute if there are updates, so messages are often delayed.
  • Messages from private Twitter accounts are not visible in the RSS-feed.
  • You can only read Twitter messages, you cannot send Twitter messages.
  • This is only possible if direct internet access to these web based feedreaders is not blocked.

Option 4 : Twitter Digest : Twitter Digest generates a daily digest of all Twitter messages of the Twitter accounts you supplied, either on a web page or thru an RSS feed
Disadvantages
  • You have to enter manually all twitter accounts in Twitter Digest you are following and you have to keep this list in sync with the actual list on the Twitter platform.
  • Messages from private Twitter accounts are not visible in the RSS-feed.
  • You can only read Twitter messages, you cannot send Twitter messages.
  • This is only possible if direct internet access to Twitter Digest is not blocked.

Option 5 : There are several sites that allow you to post Twitter message on your account by sending them an email
Examples of these services are EmailTwitter, Mail2Twitter and TwitterMail.
Disadvantages
  • In quite of lot of companies the email system automatically adds a footer to each outgoing email message. Would you like this footer to appear on each of your Twitter messages ?
  • Do you trust these services by providing them with your Twitter id and password, either by sending these by email with each message you want to post (Mail2Twitter) or by registering on their website (TwitterMail) ?
  • Does the email policy at your employer allows you to send this kind of messages thru your corporate email account ?

If your employer has chosen to block Twitter's website, I am not sure that your employer will appreciate that you apply some or all tips described in this post. If you really want to stay in touch with your Twitter friends during office hours, look for another employer.

October 10, 2007

Twitter Number of Users

TwitDir has identified until now (October 10, 2007) more than 500,000 Twitter users. TwitDir is only capable of detecting Twitter users with a public profile, meaning that 500,000 is an underestimation of the total number of Twitter users.


In July 2007 I reported that the number of Twitter users detected by Twitdir grew by 2,000 new accounts per day. Looking at the numbers reported by Twitdir recently I have the impression that this growth has slowed down to 1,500 new users per day.

October 7, 2007

First state of the Twitosphere in Chile

Last week I published an analysis of the Twitosphere in Argentina. In August I published an analysis of the Twitosphere in Brazil. On this analysis (for Brazil) a commenter mentioned an Argentinian initiative called Tuitiar.com but he also indicated that Chile has the highest Internet penetration of Latin America. I wanted to verify wether Chile has also a very active Twitter community.

Using the same approach as for all previous analyzed countries, I was able to find 593 active Twitter accounts from Chile. A Twitter account was considered to be held by someone from Chile if the location explicitely refers to Chile or if the Twitter user was located in Chile according to one of the Twitter mapping mashups. The actual number of Chilean Twitter users is probably higher as only 50% of the Twitter users specify their location in their profile (see this previous post).

For a population of 16.6 million people (according to Wikipedia), Chile has 593 Twitter accounts, which means that there is a Chilean Twitter account for every 28,000 inhabitants. Currently for Iceland and the Netherlands there is one Twitter account for roughly 10,000 inhabitants. The two other South American countries had a significant lower penetration of Twitter (Argentina : 1 for every 45,000 inhabitants, Brazil : 1 for every 451,000 inhabitants).


History of the Chilean Twitter accounts
This graph below shows the history of the Chilean Twitter accounts. The graph is based on the date of the first message posted on each of the public Chilean Twitter accounts.



Oldest Twitter accounts from Chile
http://twitter.com/OjoPiojo - Rodrigo Guaiquil - first message on 12/10/2006
http://twitter.com/paulbeelen - Paul Beelen - first message on 22/10/2006
http://twitter.com/ochovio - ochovio - first message on 17/11/2006
http://twitter.com/juque - Juan Pablo Aqueveque - first message on 17/11/2006
http://twitter.com/wachunei - wachunei - first message on 18/11/2006

The first Twitter user I was able to identify only showed up mid October 2006, several months later compared to other countries. The popularity of Twitter started to grow from March/April 2007. The graph reveals that this growth is still continuing, with even an acceleration from August 2007.


Public of Private
77 Twitter acounts from Chile or 13% have chosen to keep their updates only available to their friends. In the analyzed European countries this percentage varied from 3.5% (Italy) to 22% (Belgium). For Argentina this percentage was only 3.2%.


Number of following
A Twitter user can choose to follow one or more other Twitter accounts, called "following". 26% of the Chilean Twitter accounts do not follow other Twitter users. An average Twitter account from Chile follows 14.4 Twitter accounts. In the analyzed European countries this average varied from 9.6 (Belgium) to 32 (Italy). For Argentina the average was much higher : 46.8. Chilean Twitter accounts are on average followed by less other Twitter users compared to Argentinian Twitter accounts.



Top 5
http://twitter.com/crhistian - crhistian - 174 following
http://twitter.com/fmeza - Fernando Meza - 162 following
http://twitter.com/DonPepe - Don Pepe - 157 following
http://twitter.com/dubo - César Dubó - 153 following
http://twitter.com/eduardoe - Eduardo Escarez - 146 following


Number of followers
A follower is someone who has indicated another Twitter account as following and receives all updates of that other Twitter account. 9% of the accounts have no Twitter followers at all, meaning that no one else on Twitter has indicated this account as following. An average Twitter account from Chile has 16.5 followers. This average is in line with the averages observed for the European countries : between 10.5 (Belgium) to 26 (Italy). The average for Chile is however again lower than the average for Argentina : 46.



Top 5
http://twitter.com/mariapastora - M. Pastora Sandoval - 210 followers
http://twitter.com/dubo - César Dubó - 196 followers
http://twitter.com/OjoPiojo - Rodrigo Guaiquil - 173 followers
http://twitter.com/crhistian - crhistian - 162 followers
http://twitter.com/Paloma - Paloma Baytelman - 153 followers


Updates
5% of the accounts from Chile are still waiting for their first update to be published. Currently an average Twitter account from Chile has 215 updates, which is again higher than the averages observed for the European countries from 72 (Belgium) to 106 (Italy and Spain). This time the average for Chile is comparable to the average for Argentina (225). Please keep in mind that these averages are only snapshots, as lots of new Twitter messages (or Tweets) are published every day.



Top 5
http://twitter.com/sepulveda - Diego Sepulveda - 4.939 updates
http://twitter.com/latercera - La Tercera - 4.917 updates
http://twitter.com/Cooperativa - Radio Cooperativa - 4.313 updates
http://twitter.com/Sauce - Sauce - 3.879 updates
http://twitter.com/rots - Rodrigo Vera - 3.832 updates


Degree of activity
37% of all Twitter accounts from Chile with a public feed did not post an update in the last 20 days. They can perhaps be considered as inactive accounts.

48% of all Twitter accounts from Chile with a public feed did post at least one message in the last 7 days, indicating that these accounts are still alive. This percentage (48%) is comparable to the percentages for Argentina (45%) and Brazil (52%). It is significantly higher than the percentages of active users observed for the European countries. The South American Twitospheres seem more active than the European Twitospheres.


Conclusion
This state of the Chilean Twitosphere is only a snapshot, meaning that the figures mentioned in this post have already changed because the Twitosphere is a very dynamic environment with several new accounts being created every day, new connections being made between accounts (following/followers) and of course new messages being written every hour of the day.

Users from Chile were rather late to start using Twitter, the first Twitter user only appeared mid October 2006. The popularity of Twitter in Chile started to rize from March/April 2007, about at the same time as for the analyzed European countries. Where the growth in Europe slowed down, the creation of new Twitter accounts in Chile continued steadily.

The Twitter community in Chile is in absolute numbers not as big as the Twitter community in Argentina, but the number of Twitter users divided by the size of the population indicate that there are relatively more Twitter users in Chile (more than twice as much). It is very well possible that this is caused by the fact that Chile has the highest Internet penetration of Latin America.

The main difference of the Twitter communities in Argentina and Chile is that the Twitter community in Argentina is much more interlinked. On average each Twitter account is following more Twitter users.

October 4, 2007

First state of the Twitosphere in Argentina

I already published several blogposts on this blog with an analysis of the Twitosphere in several countries (see the sidebar). On the previous analysis (for Brazil) a commenter mentioned an Argentinian initiative called Tuitiar.com (a blog and a Twitter account). So it won't be a surprise that the next country analyzed is Argentina.

Using the same approach as for the previous countries, I managed to find 896 active Twitter accounts from Argentina. A Twitter account was considered to be held by someone from Argentina if the location explicitely refers to Argentina or if the Twitter user was located in Argentina according to one of the Twitter mapping mashups. The actual number of Argentinian Twitter users is probably higher as only 50% of the Twitter users specify their location in their profile (see this previous post).

For a population of 40,3 million people (according to Wikipedia), Argentina has 896 Twitter accounts, which means that there is a Argentinian Twitter account for every 45,000 inhabitants. Currently for Iceland and the Netherlands there is one Twitter account for roughly 10,000 inhabitants.


History of the Argentinian Twitter accounts
This graph below shows the history of the Argentinian Twitter accounts. The graph is based on the date of the first message posted on each of the public Argentinian Twitter accounts.



Oldest Twitter accounts from Argentina
http://twitter.com/Leech - Leandro Ardissone - first message on 28/09/2006
http://twitter.com/2Curious - 2Curious - first message on 29/10/2006
http://twitter.com/pugvichka - pugvichka - first message on 4/11/2006
http://twitter.com/dfgonzalez - Diego - first message on 21/11/2006
http://twitter.com/slaff - slaff - first message on 1/12/2006


Public of Private
Only 37 Twitter acounts from Argentina or 3.2% have chosen to keep their updates only available to their friends. In the analyzed European countries this percentage varied from 3.5% (Italy) to 22% (Belgium).

Number of following
A Twitter user can choose to follow one or more other Twitter accounts, called "following". 20% of the Argentinian Twitter accounts do not follow other Twitter users. An average Twitter account from Argentina follows 46.8 Twitter accounts. In the analyzed European countries this average varied from 9.6 (Belgium) to 32 (Italy). There are several Twitter accounts from Argentina with a relatively high number of following, see the top 5 below.



Top 5
http://twitter.com/Pachoro - Pachoro - 1111 following
http://twitter.com/espina - niñoespina - 1010 following
http://twitter.com/simonapalermo - simonapalermo - 800 following
http://twitter.com/elpeor - elpeor - 715 following
http://twitter.com/flowmi - flowmi - 684 following


Number of followers
A follower is someone who has indicated another Twitter account as following and receives all updates of that other Twitter account. 8% of the accounts have no Twitter followers at all, meaning that no one else on Twitter has indicated this account as following. An average Twitter account from Argentina has 46.3 followers. This average is higher than the averages observed for the European countries : between 10.5 (Belgium) to 26 (Italy). There are several Twitter accounts from Argentina with a relatively high number of followers, see the top 5 below.



Top 5
http://twitter.com/tuitiar - Tuitiar.com - 628 followers
http://twitter.com/Conz - Conz - 516 followers
http://twitter.com/espina - niñoespina - 495 followers
http://twitter.com/dariogallo - d.g. - 472 followers
http://twitter.com/mancini - Pablo Mancini - 410 followers


Updates
Only 3% of the accounts from Argentina are still waiting for their first update to be published. Currently an average Twitter account from Argentina has 225 updates, which is again higher than the averages observed for the European countries from 72 (Belgium) to 106 (Italy and Spain). Please keep in mind that these averages are only snapshots, as lots of new Twitter messages (or Tweets) are published every day.



Top 5
http://twitter.com/lanacioncom - La Nacion - 6418 updates
http://twitter.com/pabloaltclas - paltclas - 5210 updates
http://twitter.com/mtorchiari - Marina Torchiari - 5162 updates
http://twitter.com/espina - niñoespina - 5113 updates
http://twitter.com/clarincom - clarincom - 4087 updates


Degree of activity
40% of all Twitter accounts from Argentina with a public feed did not post an update in the last 20 days. They can perhaps be considered as inactive accounts.

45% of all Twitter accounts from Argentina with a public feed did post at least one message in the last 7 days, indicating that these accounts are still alive. This percentage (45%) is significantly higher than the percentages of active users observed for the European countries. This percentage is comparable to the percentage observed for the twitosphere in Brazil (52%0.


Conclusion
This state of the Argentinian Twitosphere is only a snapshot, meaning that the figures mentioned in this post have already changed because the Twitosphere is a very dynamic environment with several new accounts being created every day, new connections being made between accounts (following/followers) and of course new messages being written every hour of the day.

Users from Argentina were not among the first to start using Twitter, the first Twitter user only appeared at the end of September 2006, followed by the next one exactly one month later. There was an increase in the number of Twitter users in Argentina from March/April 2007, about at the same time as for the analyzed European countries. Where the growth in Europe slowed down, the creation of new Twitter accounts in Argentina accelerated from the beginning of August 2007.

The current Twitter community in Argentina is increasing day by day, the percentage of active Twitter accounts is higher than in the analyzed European countries. The future for Twitter in Argentina looks bright.