Map of US Congress twitter folk

This is a map of the current US congressmen and women who are currently on Twitter (you can click it to see a bigger map where you can read the names.) The direction of the arrows show who follows whom, and the size of the blobs indicates how “popular” a given congressperson is among their twittering peers (where “popular” means something like “is followed by many of their peers.”) Colours indicate party affiliation (for those of you who — like me — don’t live in the ‘States and who — like me — need reminding from time to time, the Democrats are the blue dots.)

Network of US Congress twitterers showing "citation frequency"
Network of US Congress twitterers showing citation frequency. Click for bigger.

A cursory glance at this map shows a few things:

  1. There are more Republican than Democratic congresspeople on Twitter. This is somewhat surprising because (post-Obamamania) we tend to believe that Twitter is a predominantly democratic environment (young, hip, technology friendly, coastal.)
  2. The two groups are pretty clique-y. If it weren’t for Neil Abercrombie and John Culberson, there’s little chance that information would flow between the two groups. Of course this isn’t a true representation — there may be twitterers external to the group (journalists say, political commentators, analysts, lobbyists and civil servants) who also act as bridges between the factions, but the map does seem to indicate a strongly polarized population. We’ve seen this sort of thing before — notably in Linkfluence’s PresidentialWatch ’08 map of the political blogosphere. There are, it seems fewer shades of grey.
  3. Eight of the Twitterers (including Twitter’s own ice queen, Hillary Clinton) aren’t even in the game. Nothing suggests that someone’s still using the web solely as a broadcast platform than an apparent refusal to play nicely with the other kids. Here are a couple of self-explanatory slides I showed in March ’08 as part of a presentation on why I thought Obama was doing the social media thing better than Clinton.
    Hilary Clinton's friends and followers (March 6 2008)
    Hilary Clinton's friends and followers (March 6 2008)
    Barack Obama's friends and followers March 6 2008
    Barack Obama's friends and followers March 6 2008

    See also my previous post, Why doesn’t the Tory MP have twitter friends?

  4. John Culberson’s betweenness score

    Now the Neil Abercrombie/John Culberson bridges are so striking that I thought it would be worth looking at the “betweenness” scores for this group. Betweeness is a measure of how much (or little) a node controls the flow of information in the network, its importance to the overall structure. It is one of the ways that we assess influence in a network (roughly speaking, the other two numbers we look at are “popularity” and “authority”, but more of that another time.)

    So I ran the numbers and re-drew the map (much easier than it sounds) and got this:

    Network of US Congress twitterers showing "betweenness"
    Network of US Congress twitterers showing betweenness. Click for bigger.

    Gosh. Culberson is really quite important when you look at it this way. Now — as I think I’ve suggested — I know very little about US politics (although this didn’t stop me flapping my jaw about Obama/Clinton, you’ll notice). So for those of you who are like me here’s a quick social media resumé of the congressman for Texas.

    1. He blogs at Let Texans Run Texas;
    2. He’s a fairly heavy Qik user (well, compared to me);
    3. His twitter stream consists predominantly of responses to questions and suggestions sent to him by other twitterers (how refreshing!)
    4. He has in the past used Twitter and Qik to broadcast from the House and the Oval Office;
    5. In a highly polarized map, he is one of the few Republicans to attract Democratic followers among his peers on Twitter.

    I know nothing about his politics. But you have to admit, he’d be an interesting man to meet.

    How we make these maps

    We’ve built a programme that spiders any list of Twitter people you give it for their friends and followers. There are plenty of these on the web, but ours is pretty good.

    It spits out data in VNA format. Here is a zip file of the data we collected. (216 KB).

    Generally we pre-process the data. I use a perl script or two to zoom in on the data we’re looking for. I’m happy to share this should anyone want to play with it.

    We read the VNA data into UCInet and NetDraw (both can be downloaded from Analytic Technologies) for more processing and analysis and then mapping.

    Hope this helps. I’m happy to take anyone through the process in some detail should they care to know more. As ever, all suggestions and feedback gratefully received.


  1. Dan O'Connor says

    I like this. Some further thoughts:

    1) As a percentage of the no. of Twitterers in Congress, the Dems do indeed look vastly outnumbered by the GOP (by about 2-1). This, as you say, goes against the conventional wisdom about Dems and social media. However, from this map I make it that there are 44 (? my counting of small dots is bad but roughly accurate) total Twitterers in a Congress of 535 (voting) members in both houses – 14 Dems and 30 GOPs. This means that about 8% (8.22) of Congress Twitter. Of that 8%, the GOP make up 5.6% and the Dems make up 2.62% (again rough figures). Within the larger set (all of Congress) I’m not sure those small percentages are statistically significant.

    2) Another way of looking at (1) may be to think about Twitterers within parties. There are currently 178 GOP reps and senators in total. Of them, 16.85% are Twitterers. On the Dem side, there are currently 315 members of both houses, of whom 4.44% Twitter.

    So, that’s 16.85% of GOP v. 4.44% of Dems. That seems like a more telling statistic to me.

    3) The map shows how well each Twitterer is followed ‘among their peers’… it would be interesting to re-do the map with the *total* number of followers, rather than just those within Congress.

    4) I notice that of the 40 congressional Twits, only 5 are women (Clinton, Pelosi, Boxer, Candice Miller, Michelle Bachman). That’s 12.5%. I think ‘Time’ had males as about 64% of Twitterers, ie: 26% women, so that puts congressional women pretty far behind the national average.

    5) It would also be pretty cool to map this onto each Twitterer’s political score, either from the Nat.J’s ‘liberal’ score (, or TheHill’s ‘power rankings’ (, to see if political leanings or influence have any correlation with Twittering. For example, Culberson ranks 348/435 in the House. Pelosi, on the other hand, ranks 1/435.

    I have to stop now or I will geek out all evening.

  2. says

    I love your geekiness, Dan. This is brilliant (of course.)

    I can re-draw the data. Adding things like the relative political scores or power rankings is fairly easy (we can express those as scores as sizes; the software I’m using to draw the maps isn’t too good at handling colours.)

    Our tool grabs “followers” as well as “friends” (for those of you who don’t spend time on Twitter, friends are people to whom you link, while followers are people who link to you — it’s the asymmetric nature of Twitter relationships that makes them more meaningful than, say, theFacebook relationships we’ve looked at in the past.) We rarely use it, so foolishly, I threw that data away before running the tests yesterday. Fortunately, Tweetcongress gives me summary counts for all the data, so I don’t have to set the spider loose again.

    If I get time tomorrow evening, I’ll pull those network charts together and send them to you for comment.

  3. austinwheelock says

    I really like your analysis. Where it really can get interesting is at the next level. Research staffer’s twitter accounts and see the relationships between them. It might expose the hidden relationships between congressional offices.

  4. says

    @austinwheelock — interesting you should say that. In the past analyses of corporate social networks have indeed highlighted the impact that informal networks have when compared (say) the hierarchical organograms that we so often rely on.

    To anyone who has noticed that the PA mafia controls the flow of information between operational silos, this seems obvious. And yet — on a daily basis — we ignore it.

    Where do you suggest that I get hold of the staffer information I’d need to pull that map? Would be fascinating to draw.

  5. dba says

    Wow. That is a cool thing to do. Simply amazing. I am wondering how you were able to calculate the betweenness? It seems to add a lot of meaning to the graph. Thank you beforehand!!!

  6. says

    @dba — Betweenness is calculated by counting how often a node (in this case a congressperson) lies on the shortest path between two other nodes.

    On a purely practical level, we use tools called NetDraw and UCInet (available from Analytic Technologies) to do a lot of the heavy lifting.

    For more on “Betweenness”, see this article on Wikipedia. Or there’s a really good PDF download from a talk by Steve Borgatti, one of the people whose work has really affected what we’re doing.

    Hope this helps!


  1. […] Network map of US Congress twitterers | Mediaczar – Gosh. Culberson is really quite important when you look at it this way. Now — as I think I’ve suggested — I know very little about US politics (although this didn’t stop me flapping my jaw about Obama/Clinton, you’ll notice). So for those of you who are like me here’s a quick social media resumé of the congressman for Texas. […]

  2. Twenty Items of Interest (v.43)…

    1.  I think any U.S. Legislator who was for the closing of Gitmo, lost the right to say, “not in my neighborhood.”
    2.  Palin book deal rumor going around… FALSE – according to Greta Van Susteren (who seems to have an in wit…

  3. Republicans seem to rule the roost over on Twitter…

    According to data gathered by Mat Morrison on the Mediaczar blog Republican’s in congress rule the roost over on Twitter. Even though Obama has been very willing to embrace technology it seems his cohorts in Congress have not. As his post point…

  4. Measuring Social Media Conversations…

    Measuring ConversationView more presentations from Broadsight. Here are my slides from my talk about Measuring Conversations at Enterprise Social Media Event last week the notes below serve as a commentary: Its just another Channel Social Medi…

Please tell me what you think.