Creepy app maps out the collective mood of any given population using social media posts

All your social media posts leave a trail of data relating to emotional status, with many updates revealing some level of sentiment that indicates how users are feeling. A frustrated tweet from a student or a joyous message from a sports fan both say add to the collective mental state of a population.

A new Web app called GeoMood takes the tweet stream and applies sentiment analysis to show how wide swaths of the world are feeling at any given time. It's the latest iteration of something called "sentiment mapping." The algorithm rapidly parses each tweet and filters it according to the AFINN-111 wordlist, which is a collection of 2,477 English words already rated for emotion on a scale of -5 to 5.

Tweets rated positive are green, and those rated negative are red; more neutral sentiments appear as dot outlines. A map offers a granular view of each of the world's regions, with a quick filter available for the USA, UK, Australia, and Canada.

Especially intriguing is the ability to click on a specific dot to see the content of the specific tweet that was analyzed. It's a fascinatingly creepy way to invade deep into Twitter's publicly shared tweets and learn a bit about a random human in the world.

Similarly, another project looks at geo-located tweets for emotional indicators specifically related to the city of London. It's called London Feels, and it consistently seems to suggest that Londoners only ever feel average. That makes a modicum of sense, given the British sensibility towards the more muted expressions of joy.