Another month, another visualization method!
This time it’s Twitter StreamGraph, which tries to show a river of topics related to a keyword over time. It’s early days for this particular method, I think – having a range of options, such as number of tweets, or a particular time period, etc., would be very handy, but still a very promising tool. My first use is below…
The Twitter StreamGraph above shows the last 200 posts from this afternoon. Not surprisingly, given it’s our first decent snowfall in Dublin in about six or seven years (and even that one was a gone-in-a-day jobbie), very little about recession and instead lots and lots of snow!
Cheese ,
Yes it would be interesting for access to the variables. For example, I wonder how much this would differ for the tweets from Ireland, rather than the ones that mention Ireland. Another interesting thing would be to see the delta frequency between one day ago, one months ago, one year ago etc..
I imagine they are trying to limit the amount of caching they have to do in the back end but I think that through some random sampling methods they should be able to get a fairly representative dataset over a longer period of time / volume.