The deluge of tweets produced by the Vice Presidential debate were overwhelming as they cascaded down the SayZu tweet wall tonight. But SayZu’s live animated cloud distilled the flood into some essential patterns that were easy to absorb.
At the start of the debate, the SayZu cloud was dominated by mentions of “Obama”, “Romney” and “last debate”, but mentions of “Paul” and “Biden” soon outpaced these terms, with “Biden” quickly becoming prominent.
After several minutes, “RT” started to become prominent so I used the SayZu recloud feature to create a subcloud of what was being ReTweeted the most. The term “malarkey” was immediately salient in the retweets. When I clicked the word, the SayZu side panel pulled up dozens of examples of people referring to Joe Biden’s use of the “malarkey” early in the debate.
As the cloud evolved, the text paralleled the debate in near lock-step. At a certain point the words “rude” and “disruptive” became large. Clicking on them brought up hundreds of tweets which reclouded showed “Joe Biden” as the main focus of these adjectives, with various pro and anti phrases woven into the cloud.
Simultaneously the words “better”, “lively” and “spirited” started to emerge. Fleshing out these tweets with re-cloud got mentions of both candidates.
Finally as the debate ended, I created some instant filters for various permutations of “Paul won” and “Biden won”.
“Biden won” was immediately salient in the animated cloud, and clicking the phrase immediately produced a host of Tweets carrying that message.
Interestingly, the next most salient reference to performance in the cloud were words like “Tie” and “no winner”. There was no “Paul Won” to be seen for another 5-8 minutes.
After a few “Paul won” phrases started to emerge, I clicked them to discover that Romney fans were evidently doing some sort of Tweet/ReTweet counterattack with:
RT @M1TTR0MNEY: Despite Biden’s childish demeanor, polls show Paul Ryan won by a commanding margin. RT if you agree! #TeamRomney.
“Despite” using “Despite” incorrectly, the above tweet was echoed in Republican tweets 162 times, such that even the phrase “commanding margin” became visible in the cloud.
Within minutes, Obama fans pumped out a response to the pro-Romney volley. This was not retweeted quite as vigorously but still showed up.
RT @BlGBlRD: RT If you think Joe Biden won the debate. Check with a psychiatrist if you think Paul Ryan won. #VPDebate
The pro-Biden tweets contained the phrase “Ryan won” which was a reminder that while my “Ryan won” and “Biden won” search phrases were not bullet-proof measures. While they gave a general sense of the balance of comment, there were were exceptions that became apparent when you drilled down. Interpreting unstructured text always requires a sense of context.
And context is where the SayZu cloud’s strengths come into play. The evolving SayZu word cloud captures a sense of the totality of what’s being said while the clickable text interface allows for instant non-linear exploration of the context in which salient words are used. Clicking any word or phrase in the cloud brings up all the associated text while instant re-cloud allows you to drill down even further. Features like the “right-click dictionary” allow you instantly to bundle words to reflect common phrases, names or idioms.
Of course, if you need animated bar charts that track social media conversation feedback in real-time, SayZu does that too.
Even better, right- or left-clicking SayZu bar chart labels will instantly bring up all units of associated text, then allow you to delete, revise or combine categories to reflect the context. After all, doesn’t context drive everything?
Taken out of context I must seem so strange.
- Ania DiFranco
Check us out: www.sayzu.com