̽��ֱ�� of Cambridge - Desislava Hristova

Predicting gentrification through social networking data

sc604 — Tue, 12 Apr 2016 23:13:44 +0000

̽��ֱ��first network to look at the interconnected nature of people and places in large cities is not only able to quantify the social diversity of a particular place, but can also be used to predict when a neighbourhood will go through the process of gentrification, which is associated with the displacement of residents of a deprived area by an influx of a more affluent population.

̽��ֱ��researchers behind the study, led by the ̽��ֱ�� of Cambridge, will present their results today (13 April) at the 25^th International World Wide Web Conference in Montréal.

̽��ֱ��Cambridge researchers, working with colleagues from the ̽��ֱ�� of Birmingham, Queen Mary ̽��ֱ�� of London, and ̽��ֱ�� College London, used data from approximately 37,000 users and 42,000 venues in London to build a network of Foursquare places and the parallel Twitter social network of visitors, adding up to more than half a million check-ins over a ten-month period. From this data, they were able to quantify the ‘social diversity’ of various neighbourhoods and venues by distinguishing between places that bring together strangers versus those that tend to bring together friends, as well as places that attract diverse individuals as opposed to those which attract regulars.

When these social diversity metrics were correlated with wellbeing indicators for various London neighbourhoods, the researchers discovered that signs of gentrification, such as rising housing prices and lower crime rates, were the strongest in deprived areas with high social diversity. These areas had an influx of more affluent and diverse visitors, represented by social media users, and pointed to an overall improvement of their rank, according to the UK Index of Multiple Deprivation.

̽��ֱ��UK Index of Multiple Deprivation (IMD) is a statistical exercise conducted by the Department of Communities and Local Government, which measures the relative prosperity of neighbourhoods across England. ̽��ֱ��researchers compared IMD data for 2010, the year their social and place network data was gathered, with the IMD data for 2015, the most recent report.

“We’re looking at the social roles and properties of places,” said Desislava Hristova from the ̽��ֱ��’s Computer Laboratory, and the study’s lead author. “We found that the most socially cohesive and homogenous areas tend to be either very wealthy or very poor, but neighbourhoods with both high social diversity and high deprivation are the ones which are currently undergoing processes of gentrification.”

This aligns with previous research, which has found that tightly-knit communities are more resistant to changes and resources remain within the community. This suggests that affluent communities remain affluent and poor communities remain poor because they are relatively isolated.

Hristova and her co-authors found that of the 32 London boroughs, the borough of Hackney had the highest social diversity, and in 2010, had the second-highest deprivation. By 2015, it had also seen the most improvement on the IMD index, and is now an area undergoing intense gentrification, with house prices rising far above the London average, fast-decreasing crime rate and a highly diverse population.

In addition to Hackney, Tower Hamlets, Greenwich, Hammersmith and Lambeth are also boroughs with high social diversity and high deprivation in 2010, and are now undergoing the process of gentrification, with all of the positive and negative effects that come along with it.

̽��ֱ��ability to predict the gentrification of neighbourhoods could help local governments and policy-makers improve urban development plans and alleviate the negative effects of gentrification while benefitting from economic growth.

In order to measure the social diversity of a given place or neighbourhood, the researchers defined four distinct measures: brokerage, serendipity, entropy and homogeneity. Brokerage is the ability of a place to connect people who are otherwise disconnected; serendipity is the extent to which a place can induce chance encounters between its visitors; entropy is the extent to which a place is diverse with respect to visits; and homogeneity is the extent to which the visitors to a place are homogenous in their characteristics.

Within categories of places, the researchers found that some places were more likely places for friends to meet, and some were for more fleeting encounters. For example, in the food category, strangers were more likely to meet at a dumpling restaurant while friends were more likely to meet at a fried chicken restaurant. Similarly, friends were more likely to meet at a B&B, football match or strip club, while strangers were more likely to meet at a motel, art museum or gay bar.

“We understand that people who diversify their contacts socially and geographically have high social capital, but what about places?” said Hristova. “We all have a general notion of the social diversity of places and the people that visit them, but we’ve attempted to formalise this – it could even be used as a specialised local search engine.”

For instance, while there are a number of ways a tourist can find a highly-recommended restaurant in a new city, the social role that a place plays in a city is normally only known by locals through experience. “Whether a place is touristy or quiet, artsy or mainstream could be integrated into mobile system design to help newcomers or tourists feel like locals,” said Hristova.

Reference:
Desislava Hristova et al. ‘Measuring Urban Social Diversity Using Interconnected Geo-Social Networks.’ Paper presented to the International World Wide Web Conference, Montréal, 11-15 April 2016. https://www2016.ca/ontario-online-casino/.

Data from location-based social networks may be able to predict when a neighbourhood will go through the process of gentrification, by identifying areas with high social diversity and high deprivation.

We understand that people who diversify their contacts socially and geographically have high social capital, but what about places?

Desislava Hristova

MsSaraKelly

Gentrification in Progress

̽��ֱ��text in this work is licensed under a Creative Commons Attribution 4.0 International License. For image use please see separate credits above.

Yes

Licence type:

Attribution

Opinion: Here’s how tweets and check-ins can be used to spot early signs of gentrification

Anonymous — Tue, 12 Apr 2016 08:20:34 +0000

When you walk through a neighbourhood undergoing gentrification, you can sense it – the area is dominated by strange contradictions. Public spaces are populated by vagabonds and cool kids; abandoned buildings sit in disrepair next to trendy coffee shops; blocks of council housing abut glassy new developments.

Urbanists describe gentrification as a form of urban migration, where a more affluent population displaces the original, lower-income population. In statistics, gentrification appears as the lowering of crime rates, rising housing prices and changes to the mix of people who live there.

If we could only predict where gentrification is likely to strike next, we might be able to alleviate its negative impacts – such as displacement – and take advantage of its more positive effects, which include economic growth. That’s why our latest study – conducted with colleagues at the ̽��ֱ�� of Birmingham, Queen Mary ̽��ֱ�� of London, and ̽��ֱ�� College London – aimed to quantify the process of gentrification, and discover the warning signs.

Detecting urban diversity

We constructed four measures of urban social diversity using data from social media. By combining these measures with government statistics about deprivation, we were able to pinpoint a number of neighbourhoods undergoing gentrification in London.

Of course, social media is notoriously unsuitable for population studies, because of the “digital divide”: the split between people who can access the internet and those who can’t exists even within urban areas – so information from social media only captures part of the overall picture. Twitter users in particular are known to be predominantly young, affluent and living in urban areas.

But these are precisely the demographics responsible for gentrification. So, we used information from social media from 2010 and 2011 to define the “social diversity” of urban venues such as restaurants, bars, schools and parks.

Urban social diversity – in terms of population, economy and architecture – is known to be a factor in successful communities. In her famous book ̽��ֱ��Death and Life of Great American Cities, urban activist Jane Jacobs wrote that “cities differ from towns and suburbs in basic ways, and one of these is that cities are, by definition, full of strangers”.

Dropping in. David Abrahamovitch/Flickr, CC BY

In our work, we first measured the amount of strangers that a place brings together as the fraction of the social network of visitors who are connected on social media. This gave us an idea of whether a place tends to be frequented by strangers or friends. We further explored the diversity of these visitors in terms of their mobility preferences and spontaneity in choice of venues. Although we did not consider demographics or income levels, there is a known relationship between the wealth of people and the diversity of their geographical interactions.

We studied the social network of 37,000 London users of Twitter, and combined it with what we knew about their mobility patterns from geo-located Foursquare check-ins posted to their public profiles.

By studying the amount of strangers versus friends meeting at a bar, or the number of diverse versus similar individuals visiting an art gallery, we were able to quantify the overall diversity of London neighbourhoods, in terms of their visitors.

Networks are powerful representations of the relationships between people and places. Not only can we draw links between people where a relationship – such as friendship – exists between them; we can also draw connections between two places if a visitor has been to both. We can even connect the two networks, by drawing links between people in the social network who have visited specific spots in the place network.

In this way, we are able to extract the social network of a place, and the place network of a person. By the time we’d finished crunching the data, we could take stock of the range of people who had visited a specific place, and the different places visited by any individual.

When we compared the diversity of urban neighbourhoods with official government statistics on deprivation, we found that some highly deprived areas were also extremely socially diverse. In other words, there were lots of diverse social media users visiting some of London’s poorest neighbourhoods.

Diminishing deprivation

To find out what was going on, we took the newly published deprivation indices for 2015 and looked for changes in the levels of deprivation from our study period in 2011. ̽��ֱ��relationship was striking. ̽��ֱ��areas where we saw high levels of social diversity and extreme deprivation in 2011, were exactly the same areas that had experienced the greatest decreases in deprivation by 2016.

A prime example can be found in the London borough of Hackney. Anyone visiting Hackney might describe it in terms of the contradictions we mentioned before – but few of us could afford to live there today. In our study, Hackney was the highest ranking in deprivation and the highest ranking in social diversity in 2011. Between then and now, it has gone from the being the second most deprived neighbourhood in the country, to the 11th.

So, although social media may not be representative of the entire population, it can offer the key to measuring and understanding the processes of gentrification. Neither entirely good nor thoroughly bad, gentrification is a phenomenon that we should all watch out for. It will undoubtedly help to define how our cities transform in years to come.

Desislava Hristova, PhD Candidate, ̽��ֱ�� of Cambridge

This article was originally published on ̽��ֱ��Conversation. Read the original article.

̽��ֱ��opinions expressed in this article are those of the individual author(s) and do not represent the views of the ̽��ֱ�� of Cambridge.

Desislava Hristova (Computer Laboratory) discusses how data from location-based social networks can be used to predict when a neighbourhood will go through the process of gentrification.

̽��ֱ��text in this work is licensed under a Creative Commons Attribution 4.0 International License. For image use please see separate credits above.

Yes