Archive for November, 2009
IIW IX Review
Saturday, November 7th, 2009I continue to be a newbie in the Identity space, so the plenary session the first day is a good map for me as i go throughout the rest of the unconference.
Notes from the plenary session after the cut.
(more…)
People who looked at this item also looked at…
Related items
ACM SIGSPATIAL GIS 2009, from a distance
Saturday, November 7th, 2009Turning from a conference i (mostly) attended to a conference i had no excuse to attend, this week was the 17th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems (ACM SIGSPATIAL GIS 2009). Twitter didn’t reveal much discussion with a few tweets discovered with “sigspatial” and a few others referring to “ACM GIS”. No need to turn to any visualization software. I tried correlating the tweets with the session schedule and will wait to skim the papers when they arrive. (Last year they were delivered on CD.)
I’ve found one paper already, thanks to an encouraging tweet from @snittel, and it is a fabulous cross of geospatial referencing, data mining, and the social web: Twitterstand. Socially, there’s the attempt to identify persons who tweet about breaking news. Then the data mining component comes in using both a static and dynamic training corpus to distinguish news tweets from everything else. Once a tweet is identified as news, it is then added to a topic cluster (another incredibly challenging process for computers). Once a topic cluster is established they need to locate that news. They do this by extracting and locating place names (toponym recognition via natural language processing parts of speech analysis with named-entity recognition and toponym resolution via Geonames gazetteer) from the 140 character tweets and the user’s location. Then all the locations for a topic cluster are weighted and analyzed to geolocate the news cluster, then mapped.
After the cut, tweets and the abstract for the keynote. (more…)
People who looked at this item also looked at…
Related items
IIW IX as viewed from Twitter
Thursday, November 5th, 2009I’m not feeling well, but i can get to “Many Eyes,” so here are some visualizations of the whole corpus of IIW tweets.
The Wordle
The Wordle is useful to find good starting points for using a word tree visualization. (Many Eye’s “tag cloud” is similarly useful and additionally allows exploring the occurrences of use.) I started with “trust” below. You might change the starting word to “salmon” or “webfinger” to see the buzz about those newer initiatives.
Compare to IIW8’s Proceedings from this spring
The discussion about trust frameworks initiated by the Don Thibeau of the OpenID Foundation and Drummond Reed of the Info Card Foundation appears to have made a mark. (Unfortunately comparing twitter to session notes is definitely apples to oranges.)
After the cut, about the people tweeting. (more…)
People who looked at this item also looked at…
Related items
Day one of IIW IX from Twitter
Tuesday, November 3rd, 2009
Top ten authors for the day:
8 ciberch (Monica Keller) 8 fulling (Steve Fulling) 8 jtrentadams (J Trent Adams) 8 judico (judi) 8 ragavan (ragavan) 15 windley (Phil Windley) 16 ryazwinski (Rick Yazwinski) 17 mwhelm (mwhelm) 22 anilsaldhana (Anil Saldhana) 68 DrErnie (Ernest Prabhakar)
Iterating over this retrieval command,
curl http://search.twitter.com/search.atom\?q=iiw+OR+iiw9\&rpp=100\&page=1 >\ TweetsIIW9_Tues01.xml
i collected the recent #IIW (and #IIW09) tweets, as i did at the end of JCDL2009, using the same lame gawk script to clean up the raw output.
I was disappointed to find that the Twitter list API has no way to get the whole stream for the IIW9 twitter list. It seems the best i could do would be to get the most recent status of every listed account. This is rather disappointing, because the chatter on the list is relevant and interesting but does not consistently use the #iiw tag. (I’d blame Tied House libations, but #dataownership is being used….)
I’m further disappointed to find Many Eyes is not letting me log in, so i can not generate visualizations there. However, there is always excel, so one new analysis is looking at the number of mentions of a twitter handle in proportion to how many tweets they made. So, while @DrErnie tweeted a great deal about IIW, he had a ratio of 0.19 because there were only thirteen mentions of @DrErnie (including 6 replies and 1 retweet). @dataportability tweeted four times and had four retweets, so a ratio of 1, and @windley ended up with a ratio of 0.33 with five mentions. One of @dataportability’s tweets was “RT @jtrentadams: Coming out party of DataPortablity EULA/TOS work at IIW by @EliasBiz and @greenbes #iiw via @danielabarbosa”
I wish i had a good tool at hand to show the strengths of these interconnections. Excel is not it — at least, i don’t know how to make pivot tables give up this sort of network analysis. (But somebody else is addressing this analysis with Excel: http://www.codeplex.com/NodeXL/)
Other data analysis disappointments after the cut. (more…)
