A Quantitave Study suggests that Twitter is not Primarily a Social Networking Site

13 05 2010

A lot can be said about Twitter, Facebook, Linkedin and other social media. What is the best, the most useful, the most popular the most social (and has the least privacy-issues, hehe Facebook)?

You know I love Twitter. Twitter is a social networking and microblogging service that enables its users to send and read messages known as tweets. The tweets don’t exceed 140 characters, so your message must be very concise. For me Twitter is a very rich source of information and a useful networking site. But it is hard to explain that to others.

Some Most people think that individuals who twitter are just parroting others (hé this is called retweeting, guys!) or are just egocentric bores (“I eat cornflakes for dinner”).

Well, a recent quantitative study by a group of researchers at Korea’s Advanced Institute of Science and Technology suggest that they might just be right. … Or at least their data suggest Twitter may be less of a social site and more of a news site.

According to Haewoon Kwak et al this is the first quantitative Twitter study ever.

The researchers crawled the entire Twitter site and obtained 41.7 million user profiles, 1.47 billion social relations, 4,262 trending topics, and 106 million tweets. They looked at the follower-following topology, looked at the ranking by number of followers and by PageRank, analyzed the retweets and the tweets of top trending topics.

You can read the main conclusions in the power presentation below and their abstract for Proceedings of the 19th International World Wide Web (WWW) Conference, April 26-30, 2010, Raleigh NC (USA). Below the abstract you can also find links to two download files, enabling you to reanalyze the data
Going Social Now and ReadWriteWeb also give a nice overview.

What are their main conclusions:

  • Twitter is not very “social”
    • It is “I follow you”, not “lets become friends” and you don’t have to approve or follow back. Following thus means that you “just subscribe” to the tweets of that person.
    • Only 22.1% of the relationships are reciprocal, thus 77.9% of the relationships is one way, just one of two is following the other. Surprisingly, 67.6% of users on Twitter are not followed by any of the people they follow.
    • this low reciprocity is unlike all other human social networks.
  • For most tweople, Twitter is primarily a source of information, not a social networking or information dissemination platform.
    • The Majority of topics (54,3%) are headline topics
    • Few users reach a large audience directly.
    • The average path length between two people on Twitter is 4.12. This is much shorter than Stanley Milgram’s original experiment uncovering the “six degrees of separation” phenomena.
    • Any retweeted tweet is to reach an average of 1,000 users no matter what the number of followers is of the original tweet.
    • Once retweeted, a tweet gets retweeted almost instantly on next hops, signifying fast diffusion of information after the 1st retweet.

It is a beautiful study that highlights the topological characteristics of Twitter.

One word of caution. Twitter is analyzed as a whole. There are many subpopulations with their own kinetics and goals. So the majority of people may follow the news, and fans may follow a celebrity by the million, but there are (relatively) small niches on Twitter, like health and medicine (or science) that may not follow the same rules.
I daresay (guess) that more people in this niche follow each other and do use Twitter both as a source of information and as as  network for social communication.
But these small niches are outnumbered by others (news sites, CEOs, celebrities).
At least that is my hypothesis.

Who is going to test this??

Many different Twitter birds in a flock


Reference Management Software, Shut Down of 5 Google Apps and a Plane that Crashed.

18 01 2009

Reference Management software, shut down of 5 Google apps and a plane that crashed. What have they in common? Nothing, except that these three unrelated subjects all reached me via Twitter last Thursday eve.

[1] When I checked my Tweetdeck (a twitter client) I saw a huge number of tweets (twitter messages) about the crash of a plain in the Hudson river. It now appears that Twitter and Flickr broke the news 15 minutes before the mainstream media. Below is the first crash picture which was posted on Twitter from an iPhone, taken by Janis Krums from a ferry. Earlier (Twitter as a modern tam tam) I gave some other examples of Twitter as a breaking news platform.


[2] Twitter is also a useful tool for up to date information and exchange of thoughts. For instance some tweeple (people on Twitter) had been asking about free reference management software. I had retweeted (RT, resend) the message and Thursday eve DrShock (of Dr Shock MD, PhD) tweeted a very useful link to Wikipedia which compared all reference management software, which was retweeted to the Twitter community.

The wikipedia article gives a comprehensive overview of the following software: 2collab, Aigaion, BibDesk, Biblioscape, BibSonomy, Bibus, Bookends, CiteULike, Connotea, EndNote, JabRef , Papers, ProCite, Pybliographer, refbase, RefDB, Referencer, Reference Manager, RefWorks, Scholar’s Aid, Sente, Wikindx, WizFolio, Zotero.

The following tables are included: the operating system support, export and Import file formats, citation styles, reference list file formats, word processor integration, database connectivity, password “protection” and network versions.

Very useful (although not always accurate). See: http://en.wikipedia.org/wiki/Comparison_of_reference_management_software.


[3] @Symtym (of the blog Symtym) had just learned me how to use Google Notebook to clip and collect information as you surf the web, organize the notes in notebooks and publish the public notes automatically to twitter via twitterfeed. I found it real handy and gathered some material to write a post about it.

But then came the news, brought to me by @Dymphie (of Deetjes (Dutch)), that Google decided to close many services, including Notebook as well as Google Video, Catalog, Jaiku, Dodgeball) or as ReadWriteWeb says it: “Google Giveth, and Taketh Away”. (see announcement on the Google Operating System blog).


Although Google Notebook itself will remain, the active development will be stopped. Of course this was shocking for many faithful users, including me, Dr. Shock and many others (see comments here)


What are the alternatives? Soon @DrCris, author of several blogs including Applequack, tweeted on a solution soon to come: “Evernote is working on a Google notebook importer“. I heard great things about Evernote, many doctors seem to use it, so I might as well give it a try.


Diigo is also planning to make a GN importer (see here). Presumably other tools will follow soon.

Note added:

Two articles in Lifehacker give tips [1] “where to go when google notebook goes down” and [2] describe how you can import the entirety of your google notebook to ubernote (Thanks Dr.Shock.)


nl vlag NL flag“Reference Management software, shut down of 5 Google apps and a plane that crashed”. Wat heeft dit met elkaar te maken? Niets eigenlijk, behalve dat ik donderdagavond hiervan via twitter op de hoogte gesteld werd.

[1] Eerder gaf ik al voorbeelden dat twitter als een moderne tam tam werkt en vaak een primeur heeft. Donderdag was dat ook het geval. De eerste berichten van het neerstorten van een vliegtuig in de Hudson rivier kwamen via twitter binnen.

[2] Twitter is ook nuttig om informatie te delen. Deze week vroegen mensen naar gratis reference manager software. Ik twitterde dat door (RT of retweet) en donderdag kwam @DrShock (van Dr Shock MD, PhD) met een erg nuttige link naar een artikel in wikipedia. Vervolgens werd door ‘retweeten’ een groot aantal volgers op de hoogte gesteld

In het artikel wordt de volgende software vergeleken: 2collab, Aigaion, BibDesk, Biblioscape, BibSonomy, Bibus, Bookends, CiteULike, Connotea, EndNote, JabRef , Papers, ProCite, Pybliographer, refbase, RefDB, Referencer, Reference Manager, RefWorks, Scholar’s Aid, Sente, Wikindx, WizFolio, Zotero met betrekking tot de volgende punten: “the operating system support, export and Import file formats, citation styles, reference list file formats, word processor integration, database connectivity, password “protection” and network versions”.

Heel erg nuttig en overzichtelijk (in tabelvorm met kleurtjes). Zie: http://en.wikipedia.org/wiki/Comparison_of_reference_management_software.

[3] Van @Symtym (blog: symtym) had ik juist geleerd hoe ik Google Notebook kon gebruiken om teksten al surfende op het net te knippen, bewaren en verzamelen in kladbloks en vervolgens te publiceren op twitter via twitterfeed (berichten automatisch ingekort tot 140 lettertekens). Ik vond het ontzettend handig. Het is een ideale manier om snel informatie te organiseren om later te bekijken, om er een stukje over te schrijven en/of om direct met anderen te delen.

Maar toen kwam als donderslag bij heldere hemel het nieuws via @Dymphie (van Deetjes) tot mij dat uit verschillende Google applicaties de stekker zou worden getrokken. Ook uit Google Notebook. En daarnaast Google Video, Catalog, Jaiku, Dodgeball).

Google Notebook zelf zal nog wel even blijven, maar de ontwikkeling zal worden stopgezet. Natuurlijk is dit nogal een schok voor trouwe gebruikers. Eerst worden mensen geenthousiasmeerd om een nieuwe tool te gebruiken en vervolgens wordt deze hen weer ontnomen

Gelukkig twitterde @DrCris, auteur van o.a. Applequack, vrijwel direct dat Evernote werkt aan een Google notebook importeerfunctie. Ik heb erge goede dingen gehoord van Evernote en veel artsen gebruiken het, dus ik ga dat ook maar eens proberen. Diigo is ook bezig met het ontwikkelen van een GN importeerfunctie (zie hier). Waarschijnlijk zal dit wel navolging krijgen. Toch blijft het vervelend om steeds maar van tool te moeten veranderen. Maar misschien moet je dat op de koop toenemen bij gratis applicaties.

Achteraf toegevoegd

Twee artikelen in ‘Lifehacker’ gaan over dit laatste punt [1] “where to go when google notebook goes down” en [2] describe how you can import the entirety of your google notebook to ubernote (Met dank aan Dr.Shock.)

#Sciblog – a bird-eye’s view from the camera

2 09 2008

Last Saturday I learned from @AJCann and @Jobadge (Twitter) that there was a Science Blogging Conference going on in London, that you could virtually attend.

Although I planned to do something else (banking for my mom, pick up my daughter from her overnight stay; Saturday is my-shopping-&-bodyshape-sauna- & blogging-if-I-have-some-spare-time-day), I decided to follow it. In the meantime I tried to blog about something else, which didn’t work.

I largely followed Cameron Neylon’s streamed video on Mogulus. It’s main value was the audio-stream, as well as the candid-camera function peeping at the audience from behind.

I came in late (back from banking) and unfortunately missed the Keynote lecture of Ben Goldacre from Badscience.

The next session didn’t do it for me, partly because the 3 blogging ladies ( Jenny Rohn, Grrl Scientist, Anna Kushnir) were almost inaudible and what they had to say about the bridging function of blogs between scientists and the general public (also figuratively) didn’t catch my ears. In the meantime the virtual attendents including, Fang (Mike Seyfang) from Australia, AJCann, some other guys and me, chatted in Cameron Neylon’s room.

In between I followed Twitter-messages having the hashtag #sciblog (see here). I was not familiar with hashtags, but it is a predefined tag you can add to you microblogging post to easily tract what is being said about a subject (even when you don’t actually follow the persons themselves, so as a spin off you can get acquainted with some real interesting people).

Example of a twitter message on #sciblog:

#sciblog matt woods: friendfeed encourages discussion and closes feedback loop 9 minutes ago from TwitKit

However, Hashtags is an opt-in service. You must follow @hashtags -and it has to follow you- for the service to index your tweets, so it took me some time to get it done (For more information, see this twitter wiki.) Althoug the procedure in itself was very effective, the twitter messages didn’t add much value for people already attending.

Another online backchannel, the Friendfeed room appeared more lively, but I soon stopped following the threads. Furthermore I ‘m so old-fashioned that I think speakers do deserve my attention while they’re talking (but perhaps that is because I’m not yet used to chatting at the back-scene). Checking my notes afterwards with the Friendfeed comments was useful however.

Next I followed Matt Wood’s introduction to microblogging and aggregation services and Breakout 6 “Communicating Primary Research Publicly” by Heather Etchevers (Human in Science), Jean-Claude Bradley (Useful Chemistry) and Bob O’Hara (Deep Thoughts and Silliness).

I found these presentations interesting, but tracking my notes back I couldn’t see where Matt ended and the others began.

During his lively presentation with a lot of gesturing, the heavy “sequencer” Matt Wood from “Green is Good” told us he had decided not to worry to be open and just send the message out to the public. You could use blogs to communicate your scientific findings, but blogposts do not handle versioning, although you can sometimes manipulate the post’s date (WordPress blog). Another tool is microblogging services. Twitter is more of a social platform, whereas Friendfeed is more apt for more information-exchange (no 140 character-limit). A new microblogging service is identi.ca. (see for instance this readwriteweb post)

Labnote books (and wiki’s) were a recurrent subject through the 4 presentations. They are very useful to blog primary research. People should write their motives, use it as a diary (writing down all details and circumstances), recording the results (videorecording, freehand sketches, figures, prints, text), followed by periodic summing up.

Why this is useful?

  • You don’t have to remember it (people tend to forget) (although some lab-scientists don’t like to take the notebook along to the bench)
  • Archive of ideas, (to share with people in the lab, collaborators or even ‘the world’
  • (If open) some results may be available direct outside the lab, which may be very useful for cooperation and exchange of thoughts or help (why did my blot fail?-how to proceed?)
  • It may help as a bridge to the public, i.e. by showing if public money is being spent well or for direct communication of your data to the public.
  • The info is verifiable if you link to the real data
  • Science is far more efficient this way and results are revealed instantly. Why wait till everything is distilled out? The scientist’s approach is as Hans Ricke quoted Richard Feyman from his Nobel Lecture 1966 (at Bob o’Hara ‘s blog) :

“We have a habit in writing articles published in scientific journals to make the work as finished as possible, to cover up all the tracks, to not worry about the blind alleys or describe how you had the wrong idea first, and so on. So there isn’t any place to publish, in a dignified manner,what you actually did in order to get to do the work.”

As Hans said blogs may fill that hole, because they are the place to publish this!

Major Pitfall may be that journals may not accept data reported on a wiki. And another that some people may run away with your ideas. By writing it all down you make it very easy on them. Still if everybody would become open…. For Science that would be a great good.

What I liked most of these presentations is the openness and the creativity of the presenters.
As a (medical) librarian and a scientist these thoughts came to my mind:

  • I’m a bit jealous that I worked as a scientist in the web 1.0 era. This way of approaching science looks very stimulating to me, but maybe that’s only a romantic look from the outside?
  • How do we as librarians step in? Can we play a facilitating role? Should these primary findings be aggregated and made available in a searchable way?
    We should at least keep more in pace with the new scientific developments and the way researchers exchange and find their information. It’s entirely different to what we are used to. (we= most librarians I know, including myself)
  • I wonder if such an approach could also be used in medicine and/or in EBM. Are wiki’s like this useful for CATs for instance? Question, PICO + domain, best study type, search, critical appraisal, summary, power point presentation, pdf-files, video of CAT etc??? link to video of casus perhaps?

To get an impression of the great features of such a wiki/open notebook, take a look at http://usefulchem.wikispaces.com/ (Jean Claude Bradley). You can also go to the Useful Chemistry blog and click at “UsefulChem wiki”. Note for instance the links to the notebooks of the individual scientists. Really impressive.

Below you also find the (short) presentation of Heather. Hope the others will follow soon and share their presentations

more about “Sciblog2008 Etchevers“, posted with vodpod
Other Info

Twitter Traumas: Twitter’s Janus Face

7 08 2008

In a few posts I praised Twitter, the free microblogging service, for its value as a rich source of social contacts, news and ideas. See for instance this post about Twitter as a modern tamtam or this one titled: “Forget Hyves go Twitter”.

In the short period I used it (2 months) I also noticed some drawbacks: its frequent down times, for instance. The sudden disappearance of half of my followers, a phenomenon, which appeared to affect half of the Twitter community last week. The vanishing of part of the archive (@Deeboeks). However, last week the situation has come to a head by the banning of some if its most active users. Why? Because these highly following and followed twitterati were apparently suspected as spammers. Without thorough verification, that is. These addicted Twitter-users were greatly inconvenienced.

@davedelaney: “(from his blog) Take my email analogy and consider how you would feel, or as if someone took your personal journal and address book and refused to return them.”

@pfanderson: (from her blog) “Now, two important points. First point, earlier this week I stated in this blog that Twitter is my #2 productivity tool. In other words, this is REALLY important for me! Second point, Twitter funkiness (like Second Life funkiness) is not unusual, so at first I did not realize this was anything beyond the typical. (…..) All kinds of alarms went off. I had just been asked to demo Twitter (among other social techs) at an important upcoming meeting. Wait a week? For them to just look at the problem? Ummm, that could be a REAL problem!”

Well, it isn’t bad when spam is banned. Indeed many of the spamming or advertising twitterati are following a huge number of people, but are followed by relatively few. Thus a ratio of 1500:50 sets the alarm bells ringing. But then you must do a second check: of profiles, website links and tweets, of course. A proper check of @davedelaney, @tibbon, @conniecrosby, @skalik, @marjarpanic, @abrudtkuhl, @pfanderson (one of my favorites) @narain (a twitterer I referred to as bringing breaking news about Bomb blasts in Bangalore!!) would have learned that these people intensively use twitter for serious purposes. Quite different from other twitterers that follow a whole crowd, only producing twitter messages like: “want to do X than look at this (= my) website” or all kind of quack. And these twitter-spammers are not banned…

But everybody can make mistakes, as long as you….. right, 3rd problem….. communicate it directly to your customers, preferably before you ban them. And if you faile to do so, … right 4th mistake…. do you best to fix the problem and …..o.k. 5th really unforgivable mistake …. always give yours sincere apologies!!

Nothing of the kind happened and that is really bad. If I could, I would ban @Ev (Twitter’s silent leader who was enjoying wine and pie while some of his most loyal users were panicking, according to @davedelaney‘s description) at least for a while. Without checking, without notification, without responding to his frequent requests for help, and foremost without any apology afterwards!!

As far as I know all accounts are back now, but it has cost these people a lot of precious time and has changed their feelings about twitter as a program.

@davedelaney has a poll at the end of one of his post asking:

Are you having second thoughts about Twitter now?”

66% of the responders gives an affirmative answer.

There are some Twitter alternatives, but as Delaney says:

“Of course my main love for Twitter is the Community (…). Without the people there would be no Twitter. I don’t love Twitter, I love the people on Twitter who make it such a vibrant place to be. I don’t love Twitter, I don’t even like Twitter at the moment – I may even hate them.”

Further reading:

On Dave Delaney’s blog (August 3rd) : 5-reasons-why-i-hate-twitter. Very balanced: he also gives 5-6 reasons why he loves Twitter; some alternatives mentioned as well)

An account of the banning adventure :can also be found at the blogpost of @pf anderson: twitter_banning.html and on here slideshow on Flickr


Eerder promootte ik Twitter als een microblogging-dienst, omdat het zo’n rijke bron is van sociale contacten, nieuws en ideeen. Zie bijvoorbeeld de berichten: Twitter as a modern tamtam en “Forget Hyves go Twitter”.

Maar in de korte periode dat ik het gebruik, ondervond ik ook enkele nadelen. Twitter is bijvoorbeeld herhaaldelijk niet bereikbaar (“down” , “upgrading” of te zeer belast). Vorige week verdween ongeveer de helft van mijn fans (en ik heb er al zoveel), maar “gelukkig” bleek ik niet de enige te zijn. Soms verdwijnt een deel van het archief, bij @Deeboeks bijvoorbeeld: “Al mijn tweets van voor 6 april 2008 zijn uit ‘t archief verdwenen”. Maar vorige week was toch wel de klap op de vuurpijl. Opeens waren de accounts van de meest actieve twitteraars verdwenen. Waarom? Ze werden kennelijk van spammen verdacht en zonder staving verwijderd. Dat bracht behoorlijk wat ongemak met zich mee, daar dit juist mensen zijn die Twitter veel gebruiken en er deels ook afhankelijk van zijn. Voor @pfanderson: is Twitter is haar “#2 productivity tool“. @davedelaney: “zegt het als volgt:

“consider how you would feel, or as if someone took your personal journal and address book and refused to return them.”

Natuurlijk is het bannen van spam niet slecht. Als de verhouding following/followers 1500:50 dan moet er wel een belletje gaan rinkelen. Maar dan moet je vervolgens het e.e.a. checken: profiel, evt. website en natuurlijk de aard van de tweets.
, @tibbon, @conniecrosby, @skalik, @marjarpanic, @abrudtkuhl, @pfanderson (een van mijn favorieten) @narain (die het nieuws over de aanslagen in Bangalore bracht!!) zijn intensieve en serieuze gebruikers. Dat kun je in één opslag zien.

Iedereen kan fouten maken, maar je moet natuurlijk wel goed en tijdig communiceren met je gebruikers, alles doen om de fout te herstellen en vooral achteraf je excuses aanbieden. En dat is allemaal niet gebeurd.

Wat zou ik graag Twitterbaas @Ev (die alleen twitterde dat hij heerlijk genoot van wijn en quiche terwijl zijn klanten in paniek waren en om zijn hulp vroegen) het zelfde lot doen ondergaan: account rucksichtlos verwijderen, niet op de hoogte brengen, niet reageren en niet helpen. En natuurlijk al helemaal geen excuses aanbieden!!

Ik geloof wel dat alle accounts nu weer in orde zijn, maar het heeft deze mensen wel veel ergernis bezorgd en hun gedachten over Twitter veranderd.

@davedelaney heeft zelfs een poll gezet onder zijn Twitterbericht met de vraag:

Are you having second thoughts about Twitter now?”

66% van de mensen die hierop gereageerd hebben zegt ja.


Get every new post delivered to your Inbox.

Join 607 other followers