Tuesday, July 04, 2006

Social network of 911

Cooperative Research is a good resource for finding organized timelines for various important current events topics (e.g. Katrina, 911). In particular, the timelines are available in XML and easily parsed.

I was able to extract the timeline for 911 and, for each event in the timeline, the players involved in the event. Using the same network tools I've been using the other blog postings I've done recently, I have plotted the social network of the major entities involved.

For each picture, you can get a larger version by clicking on it. Note that a link means only that the two entities were involved in the same event(s), not that they necessarily acted together. Edge color represents the number of separate events in which the attached nodes were both involved, and vertex color represents the closeness score (for the first plot), the betweenness score (for plots 2 & 3) or the number of links (plots 4-6) for the entity.

Entities with top 75 closeness scores:

Entities with top 50 betweenness scores:

Entities with top 100 betweenness scores:

Entities with > 750 links:

Entities with > 500 links:

Entities with > 250 links:

Labels:

6 Comments:

Anonymous Anonymous said...

Luv luv luv the website, very stimulating...

What perchance do you guys know about The statistically improbable phrase's or SIP's?

Further, what about, sip chains.?

And even further, and more importantly, isn't something like
this

the reason why and or how the NSA, and Bush wants to profile everyone.
by using network algorithms?

10:05 PM  
Anonymous Anonymous said...

addendum, A quote from MSNBC might illustrate my muddled thinking on this topic-

"Search is a window into people's personalities," says Kurt Opsahl, an Electronic Frontier Foundation attorney. "They should be able to take advantage of the Internet without worrying about Big Brother looking over their shoulders."

Also, EFF's (Electronic Frontier Foundation's )Kurt Opsahl argued the case in a podcast
recently, ..good listening

10:12 PM  
Blogger slomo said...

Re: SIP's, I don't know a whole lot about them except that I gather they are based on Markov-chain models of language: for any two words w1 and w2, there is an estimable probability of w2 occuring after w1, i.e. P(W[n+1]=w2|W[n]=w1). Using the laws of conditional probability, you can construct the probability of any phrase.

Re: political implications. Exactly. That's why I'm bringing all this up. I don't pretend that a single blogger with the computing power of a home PC (or a university cluster used surreptitiously) can outmatch the US gov. But I think we can at least try to turn the tools of our surveillance society around and direct them back at those doing the surveilling!

8:48 AM  
Anonymous Anonymous said...

interesting; $cam receives his wish

don't have time at the moment slomo to fully explore & digest your site; what i've parsed so far - fascinating

tx


....

6:53 PM  
Blogger Interrobang said...

Hi, got here from No Capital, will be back. That is a very interesting analysis. I also went over and had a look at your Social Networks article as well.

I'm currently trying to build something similar mapping the relationships between a certain very large corporation, its subsidiaries and strategic partners, and governments and policymakers around the world. I already have a huge set of data points, but I could use some methodological advice. If you could give me some, I'd really appreciate it. My e-mail is shgstewart at gmail dot com.

Thanks.

1:27 AM  
Blogger خدمات منزلية said...

شركة تنظيف فنادق الاحساء
شركة تنظيف بلاط بالاحساء
شركة تنظيف سيراميك بالاحساء

7:10 PM  

Post a Comment

<< Home