Timur Abbiasov

Do Urban Parks Promote
Racial Diversity?

Evidence from New York City

In my research paper [link], I examine individual GPS tracking data of more than 60 thousand Twitter users in New York to find out whether the city's parks help to make the casual encounters of its residents more racially diverse.

How do cities integrate their communities and promote racial tolerance? Scholars and policymakers argue that parks and public spaces can help.

By coming together in these relaxed settings [parks], different groups become comfortable with each other, thereby strengthening their communities and the democratic fabric of society...
Setha M. Low

Parks and public space are essential to [...] promoting interaction
Mayor Bill de Blasio

Quantifying the Everyday Encounters Using GPS Data from Social Media

How can we measure indivial exposure to racial diversity on a large scale? One solution is to harvest the vast geospatial data posted on social media to estimate the chances of casual encounters between different people.

I use a unique dataset of more than 15 million geotagged Twitter posts published between June and December 2014 and originating within the New York metro area.

Overall, I obtain high-resolution GPS travel data for more than 60 thousand city residents.

Then, I apply commerical machine learning to process the profile images and obtain the perceived ethic or racial attributes.

Combined with the information on residence locations — also inferred from the patterns of online activity — this approach allows me to create a detailed representation of how different racial or ethnic communities co-locate daily in urban space.

To measure individual experienced diversity, I map the set of locations visited by each indivdual into a grid of rectangular cells (approx. 150 by 150 meters).

Then, for every location, I obtain the representative racial and ethnic composition of users that visited the same geographic cell in a particular month.

Finally, using the monthly frequencies of observed visits to each cell, for each user I estimate the experienced diversity index as the probability of encountering a person of different race or ethnicity on a random trip from home.

The central aim of this project is to establish whether the availability of park space has an effect on individual diversity.

In a cross-section setting, residential selection presents a major threat for causal inference. For example, if more sociable people value parks more highly, one would likely find that users with more parks in their neighborhood are more diverse, but this does not imply that parks increase diversity.

In my paper, I circumvent the issues arising from static selection by incorporating the data on temporal park service disruptions due to ongoing contruction works. In particular, I construct a measure of park access that varies with time for each individual. This approach allows me to estimate the causal effects using a panel regression with individual fixed effects.

My results show that for White and Black/African American residents additional 10 acres of park space within a 5 km radius from home increase individual chances of encounters with the members of other groups by 1 percentage point.

Conclusion

The present work emphasizes the role of parks – and, more generally, of public space – in nurturing diverse social environments and suggests a viable policy that can help cities reduce racial isolation.

[link]