Studying emoji usage through geolocation.

Andalucía, the southernmost community in peninsular Spain, just celebrated its regional holiday. Every 28th of February, students all over Andalucía gather in their playgrounds and eat the much-celebrated “panaceite”: a slice of bread with locally made olive oil, as well as stuff like chocolate, sugar or green beans. People wish each other a happy day on social media, but there’s a (very unimportant) problem: the Andalusian flag does not exist in emoji form. The flag, which consists of three green-white-green horizontal stripes, looks really similar to the Nigerian flag, the only relevant difference being the vertical disposition of the latter. But that does not seem to be a problem for most.

So I was wondering: what would a heat map of the usage of the 🇳🇬 emoji look like? Is the online Andalusian frenzy enough for a trend to be noticeable? I set to find out, and the results have been interesting. The code can be found here. Some of the things I figured out:

• Very few people geo-tag their tweets. In my sample of ~ 32.000 tweets, only about 7% had coordinates attached.
• Most tweets come from Lagos, the most populated city in Nigeria:
• Andalusians do leave a trace, according to this script. They’re also the main source of tweets in Spain:
• There is a local football team from a Northern town with a population of 9.000 whose flag is very similar to that of Nigeria. Took me a while to figure out why there was a small focus in the north of the country, seemingly in the middle of nowhere. The focus is not pictured in any of the images, though.