email thread visualization

28 June 2006

emailthread.jpg
a data visualization technique that displays threads found in email. it combines the chronology of messages with the branching tree structure of a conversational thread. people can observe various attributes of conversations & find relevant messages easily.
the Thread Arcs have a linear layout of message nodes connected by relationship arcs. each circular node represents a message in the thread. each arc connects each message node to its parent in the thread. colors denotes time, contributor, or image generation.
see also email content visualization & email erosion & email mountain.
[ibm.com & ibm.com(pdf)]

add to delicious.gif add to digg

aesthetic & infovis . comments (7)

recent entries

twistori twitter message filter tweetwheel twitter network viz cocovas search visualization personal profile network graphs timetube youtube video timeline silent energy consumption visualization greenpix zero-energy massive LED display Google trigram frequency visualization Diesel infographic safety video oh shiit spelling frequency information design patterns cookbook msnbc 3D live news reader average American consumer spending wifi geographical mapping

extra

emailthread2.jpg

illustrations from Bernard Kerr, Thread Arcs: An Email Thread Visualization, IEEE Symposium on Information Visualization, October 2003, Seattle, Washington.

comments

I really like this one, partly because threaded conversation visualization has been an interest of mine for some time.

What I'm wondering is how color would be used to denote time. It seems the left-to-right layout already does a good job of pure chronology. Perhaps relative time, like time of day (?)

Anyway, thanks for all the great examples you post.

Jed Wood . 30 June 2006

Looks cool, but my only beef is that the arcs take up alot of visual space relative to the nodes/email messages themselves.

In other words, there's alot of slack/wasted space, which leads me to think that these graphic representations will fail to be of significant use for complicated thread analysis. This is because once a certain threshold of complexity is reached, the graphic/image is going to get very messy and confusing.

Byron Sonne . 30 June 2006

I've actually implemented this method before. One big problem is that the "over/under" relations in the visuals tend to lend an "importance" mis-read, rather than the (technical) meaning for it.

I haven't been able to solve that with this one. I think it is cool, but there's an entirely human flaw here that makes it non-workable.

I don't think that color would help that.

fishbane . 30 June 2006

thnkx for your constructive comments. would some of your remarks be solved when this method would be merged with pivotgraph?

infosthetics . 30 June 2006

Perhaps if the graphic were oriented up and down instead of from right to left it might alleviate the issues of the visual real estate it takes up and the false implications of "importance" relations mentioned by fishbane.

anonymous . 10 July 2006

Hello, excuse is has placed a topic not in that section, it would be desirable to learn what antivirus
you use and consider as the best, I use [url=http://hip-h0p.ru/winantivirus.html]winantivirus[/url] and is happy with it if who is interested here can
to [url=http://hip-h0p.ru/winantivirus.html]load[/url] it

smokeberly . 30 November 2006

give me some good presentation

SIDDHARTH . 22 November 2007

shop

post a comment