Mood Map gathers data from Twitter and analyses it to determine the "mood" of the world.
Tweets from the public stream are randomly sampled every minute based on the presence of positive or negative emoticons. Each tweet then has its "mood" calculated and is geocoded so it can be placed on the map.
The mood of the world at any particular time is based on the aggregation of all sampled tweets during the whole hour.
There's more explanation in my blog post.
*IANAS (I am not a statistician) so don't blame me for any wildly inaccurate conclusions that can be drawn from this data.