Tuesday, 5 February 2013

Twitter usage in London: case study

Five of data from London showing that tweeting is highly governed at the city level
The graph above is a random five days showing the levels of twitter intensity for a square KM around Bank, the financial centre, Brick Lane, a popular social spot on the east end, and Oxford Street, a major global shopping area.

As you might anticipate the rates of tweeting int these areas are very different.  Oxford streets is a massive crowed areas and its it almost always the sigh of the highest tweeting.  Brick Lane is sometimes as active as Bank Street, the site of the Bank of England and many large banks, but other times it is very low.

But what is really striking how the city level impacts tweeting.  Tweets in all three major location, despite being very different kinds of places used for different kind of things, the rates of tweeting in all locations still rises and falls in line with each other.  If you want to predict the level of tweeting in Brick Lane the best basis of prediction would be to look at the tweeting level in Bank or Oxford Street.

Now a large part of this is because of overlap.  Because of the way geotagging work the same tweet can be count as being with 1 KM of Brick Lane and Bank Street.  But there seems to be a large degree where tweeting levels are governed on a city level.

You might respond that tweets seems to be governed by time.  This does normally seem to the the fact, but the way time impacts tweeting is governed on a city wide level. The chart bellow shows a a different period with the tweets from Brick Lane in red and Bank Street in Blue.  There is a drop around Christmas and you can see both go down at the same time.

Brick Lane levels of tweeting in red and Bank Street in blue, you can clearly see dramatic rise and fall between the two of them.

Above with see the end of 2012 for midtown Manhattan in New York, with blue showing Christmas.  New York is striking in that tweeting is higher and more consistently higher, even at night.

No comments:

Post a Comment