This viz has been getting some press lately. It takes a look at tweets by country during the course of the Euro2012 tournament. Aesthetically it is pretty cool and I like the interactivity but from an information perspective there are some shortcomings.
The primary message the steam graph is communicating is that on the 23rd, 27th and 30th there were a lot of tweets and that dipped on 20th, 25th, 26th, 29th and 30th. Presumably, the reason for the surge was the importance of the games while the dip could have been caused by the fact that no games were played. What is much less clearly communicated, and seems to be the intended goal of the viz, is which countries tweeted the most, when, and why. The reader has to interpret the size of the area to understand the number of tweets for each day. It is also not very clear where the points originate from. It would be more informative if this were a line chart on the same axis originating from zero. A user could click the country and it would highlight the line for that nation.
The smaller area charts below the steam graph break out the countries not by day but during the course of the game. Each chart indicates when a goal was scored which usually correspond to spikes in tweets. These charts are much more informative and contextualized. The animation as you move between games is nice and smooth. It would probably make more sense to have these be spark line charts on the same axis so you can compare the two countries during the course of the game. Another thing that is missing is a horizontal axis label for the match time. For example, during the England-Italy match there is a spike at the end of the game. Is this spike during the penalty kicks or the end of regulation? A simple axes label would provide even more context.
Overall there are elements of this viz that are really impressive. The animation is really polished and it looks pretty cool. But from the perspective of communicating information I imagine Stephen Few would be none too pleased.