Microsoft’s Fuse Labs has released their collection of “Social Gadgets”, a few interactive widgets that analyze Twitter Streams. Above, see their “Tag Cloud” visualization of a search for “VizWorld”.
FUSE Labs’ SocialGadgets are a set of embeddable widgets that visualize Twitter real-time data. Each gadget focuses on a given keyword and displays its volume of usage over time. Mentioned entities such as people, locations, companies and noun phrases are identified and visually represented. The gadgets are interactive, letting users explore the relationships between topics publically shared on Twitter. By focusing on patterns and trends, the gadgets can extrapolate what people find important, and provide a succinct yet effective way to look at events as they are unfolding.
Now, I’ll cut them some slack because 1) it’s new, and 2) it’s still kinda beta.. But I easily see several issues:
- What is that bottom graph? The axis goes from 0 to 1.. Is that some count of tweets? Retweets? some statistical percentage? I don’t know. A search for “vizworld” at search.twitter.com fills the entire front page and only goes back 20 hours, so I obviously have more than 1 search result.
- For a “tag” cloud, it’s pulling up more usernames than hashtags.
- Where is it getting this data? I don’t see any of those accounts in my research search history for “vizworld”. And it leaves out the most common users who retweet my stuff like ebruhwiler and tssveloso.
My only guess is that they are unable to really keep up with all the data coming out of Twitter, so they must be missing large chunks. That or their algorithms are completely hosed.
Either way, go check out their gadgets and see what your results are, and share in the comments! Is “vizworld” an isolated case?
Hey Randall – tx for checking out the gadgets.
Your hypothesis is correct. We are not indexing 100% of the Twitter firehose (although working towards it!). So because we’re only seeing a small number of Tweets that mention ‘vizworld’, the tagcloud displays entities mentioned in those posts (in your case usernames were mentioned, and only one hashtag was found – ‘#ff’). Here’s the actual tweet – http://twitter.com/RemedyEditorial/statuses/26798026718
This is totally a lab experiment. The gadgets are really interesting for looking at trends and queries that generated somewhat of a buzz on Twitter. I’m hoping to understand what (if at all) people are finding useful – which types of gadgets, what type of data – and what’s enticing for them to embed on blogs or websites.
More updates will come soon! Keep you posted.
@gilgul