The gentrification of London boroughs can be predicted from tweets and Foursquare check-ins, researchers at Cambridge University have found.
By combining social media data – indicating a more affluent visitor to an area – with statistics on rising house prices and dropping crime rates, the computer science researchers were able to show a link with gentrification.
Rather unsurprisingly, Hackney was identified. Based on 2010 census data the borough had the highest social diversity but the second-highest deprivation, making it a prime candidate for gentrification. This was supported by the UK Index of Multiple Deprivation in 2015 ranking Hackney as the highest climber.
Tower Hamlets, Greenwich, Hammersmith and Lambeth are next on London's gentrification list, the researchers claimed. "There is a cluster of neighbourhoods which have very high urban social diversity according to our metrics and have extremely high deprivation," Desislava Hristova, lead author on the paper, told WIRED. These factors, she said, provided an "early warning" for an area becoming gentrified.
"Those neighbourhoods that had the highest diversity rank and highest deprivation rank were also the ones that experienced the highest improvements in the last five years," Hristova continued.
The team behind the work collected 549,797 check-ins at 42,080 London locations from Twitter and Foursquare posts. These were linked to an identifier, such as a bar or restaurant, and came from a total of 37,722 people.
The researchers identified four key factors for working out social diversity. Brokerage, the likelihood of someone visiting an area alone or with friends; serendipity, the likelihood of people visiting a place; entropy, the diversity of visitors and the homogeneity of an area and its visitors
By connecting social media users to online friends, the team was able to identify locations that tended to draw crowds of strangers or groups that knew each other. To interpret the visitor diversity of an area the team compared each user's check-ins to the different categories of places other individuals had also visited.
Hristova said that in most population studies social media data was not useful as users were more likely to be affluent and have a greater social mobility. However, for determining whether an area is becoming gentrified, it is possible to "use social media bias to advantage".
Social media data makes it possible to understand where those who are more affluent are visiting and spending their time. "The people that may be causing gentrification are precisely these social media users," Hristova said.
From this data the team – which also involved academics from the University of Birmingham, University College London, and Queen Mary University London – were able to show that the areas with the highest diversity and deprivation were likely to undergo the highest levels of change. "What becomes immediately apparent is that there is a clear distinction to be made between inner and outer boroughs in terms of diversity," the paper explained. Central London boroughs bring together more strangers and have higher levels of social diversity.
The team said that applying the diversity metrics to the data acted as "good predictors of gentrification when measured through indices of deprivation". While areas that had the most diverse visitors were likely to be going through gentrification, the researchers also said those areas that do not mix well are unlikely to change. "This suggests that affluent communities remain affluent and poor communities remain poor through isolation," the research paper concluded.
This article was originally published by WIRED UK
