This document summarizes research on using Twitter for live event annotation and discussion. Key findings include:
- Tweets during live events allow for implicit annotation through hashtags, mentions and retweets. Volume of tweets peaks during events and discussion surges after.
- Automatic analysis of tweet volumes can detect segment boundaries in events with 92% accuracy compared to editorialized summaries.
- Mentions create directed networks showing influence, while hashtags group tweets by topic. Central figures and communities form around events.
- Future work includes sentiment analysis, topic trends over time, and using tweets to understand events without source material. Tweets provide indirect annotation of and reaction to events through social discussion.
7. Social and Live Performance DJs manage three social networks through group of mediums like: MySpace, Webcasts, Twitter, Facebook, and IM.
8. Much of Social Media is about Congregation Something we think about at CHI and CSCW. (if you are so definition inclined you can enjoy the above paste)
9. A new form of indirect media-object annotation. 中国没有Twitter
13. a Tweet RT: @jowyang If you are watching the debate you’re invited to participate in #tweetdebate Here is the 411 http://tinyurl.com/3jdy67
14. Anatomy of a Tweet Repeated (retweet) content starts with RT Address other users with an @ RT: @jowyang If you are watching the debate you’re invited to participate in #tweetdebate Here is the 411 http://tinyurl.com/3jdy67 Rich Media embeds via links Tags start with #
15. Indirect Annotation Sept 26, 2009 18:23 EST RT: @jowyang If you are watching the debate you’re invited to participate in #tweetdebate Here is the 411 http://tinyurl.com/3jdy67
16. Tweet Crawl Three hashtags: #current #debate08 #tweetdebate 97 mins debate + 53 mins following = 2.5 hours total. 3,238 tweets from 1,160 people. 1,824 tweets from 647 people during the debate. 1,414 tweets from 738 people post debate. 577 @ mentions (reciprocity!) 266 mentions during the debate 311 afterwards. Low RT: 24 retweetsin total 6 during 18 afterwards.
17. Volume of Tweets by Minute Crawled from the Twitter RESTful search API.
18. Tweets During and After the Debates Conversation swells after the debate.
22. Automatic Segment Detection We use Newton’s Method to find extrema outside μ±σ to find candidate markers. Any marker that follows from the a marker on the previous minute is ignored.
23. Automatic Segment Detection with 92% Accuracy When compared to CSPAN’s editorialized Debate Summary ± 1 minute.
30. HCC and MM Findings & Future Work Indirect annotation through community action Uncollected Sources (read: events) are highly valuable Segmentation Figure Identification Term Distance What about Sentiment? Onset? Trends? Sustained Topics?
38. Thanks Chloe S., Ben C., Marc S., M. Cameron J., Ryan S.! @paulr they are coming #2 The midnight ride of Paul Revere
Hinweis der Redaktion
There is MORE to tagging and comments in social media than how we think of it currently as the single browser/site/startup.
These tags and comments are regulated to anchored explicit annotation. This is the problem. Temporally, there is a gap – we cannot leverage these components like we have with photos.
Several sites (including YouTube and my own past research) tried to make deep comments prevalent.
Enter Twitter. (explain it quickly) With twitter, when something happens and you wanna shout, you tweet.
Many People Tweet while they watch tv, many TV shows call for people to follow the twitter stream.
(this is a fake tweet)
Not only of the tweet to the video but the rich data within the tweet.
Some techniques from may be applicable: Wei Hao Lin, Alexander Haputmann: Identifying News Videos ideological viewpoint or bias