Comments on: Making sense of Twitter search

By: __fury (fury)

__fury (fury) — Sun, 23 May 2010 07:08:59 +0000

Twitter Comment

Twitter: FXPAL Blog » Blog Archive » Making sense of Twitter search: [link to post]

- Posted using Chat Catcher

By: FXPAL Blog » Blog Archive » Aggregating Twitter

FXPAL Blog » Blog Archive » Aggregating Twitter — Mon, 17 May 2010 14:51:42 +0000

[...] instead, a more holistic approach is more appropriate. I described in one such approach in Making Sense of Twitter Search (the position paper was co-authored with Miles Efron and was presented at a CHI 2010 workshop on [...]

By: FXPAL Blog » Blog Archive » Microblogging Inside and Outside the Workplace

Fri, 05 Mar 2010 15:32:07 +0000

[...] was also an interesting comment that relates to my interest in managing Twitter search results. Ehlrich and Shami write that Although BlueTwit contained mostly work related posts, possibly [...]

By: Shyam Kapur

Shyam Kapur — Thu, 18 Feb 2010 06:34:33 +0000

It looks like you guys are having a lot of fun with tweets. I am not at all surprised. I have also been working on designing some good interfaces on top of some sophisticated analytics technology for data such as tweets. I think some of you have already seen one manifestation of my work, TipTop at http://FeelTipTop.com . Visit again often to see how it is evolving. Those who have not seen it so far might want to visit TipTop soon.

By: Conference: Search and Social Media 2010

Conference: Search and Social Media 2010 — Tue, 16 Feb 2010 20:45:51 +0000

[...] For more on this and a cool demo, check out Gene Golovchinsky’s look at the SSM2010 twitter coverage. [...]

By: Making sense of Twitter Search – Gene Go… « AuthTweet

Making sense of Twitter Search – Gene Go… « AuthTweet — Fri, 12 Feb 2010 10:36:42 +0000

[...] http://palblog.fxpal.com/?p=2948 [...]

By: frostola (Tony -Frosty- Welch)

frostola (Tony -Frosty- Welch) — Thu, 11 Feb 2010 05:27:58 +0000

Twitter Comment

RT @genebecker: FXPAL: Making sense of twitter search - interesting [link to post] cc: @ristoh

- Posted using Chat Catcher

By: genebecker (Gene Becker)

genebecker (Gene Becker) — Thu, 11 Feb 2010 05:27:01 +0000

Twitter Comment

FXPAL: Making sense of twitter search - interesting [link to post] cc: @ristoh

- Posted using Chat Catcher

By: Gene Golovchinsky

Gene Golovchinsky — Wed, 10 Feb 2010 18:56:31 +0000

My guess is that temporal clustering around parts of the social graph and around hashtags might create statistically-meaningful clumps. Also would be useful to see how document content can be pushed back through the link to inform tweet grouping.

One challenge in all of this, however, is to retain transparency and predictability. If people don’t understand why some set of tweets is grouped, then it’s not useful to group them that way.

By: Miles Efron

Miles Efron — Wed, 10 Feb 2010 18:40:01 +0000

An extra challenge for those of us working with twitter data… Dealing with topic (and other aspects of meaning) is interesting/hard in this space. As you point out, most tweets are useful insofar as they form part of some group of tweets that relate to a given topic, event, or whatever. So in many cases, a tweet’s topic is latent–not directly observable from lexical features.

If we do want to enlist topical aboutness in our analysis of microblog data, I think we’ll need to do a bit of re-tooling of our standard arsenal. To take one example: measures such as TF-IDF have shown themselves to be unreliable in some initial work I’ve done… obviously TF doesn’t have much meaning here. And the intuition behind IDF doesn’t hold very well in microblogs (e.g. relatively common hashtags are often quite indicative of topic, while rare hashtags aren’t of much interest).