How do Cities Measure "Social Capital"?

By Sarah Quartey
B.A. Candidate in Urban Studies, 2014

Read about the CityNature project on the OutWest student blog. Over the summer, a team of undergraduate student researchers combined spatial analysis with innovative mining of planning document text, photographs, social media, and published historical narratives to explain why nature is unevenly distributed in and across cities.

Here’s something that wasn’t initially on my bucket list: read the comprehensive plans for each of the 40 largest cities in the United States. It’s kind of funny how things like that happen. Anyway, let’s back up and explain how I found myself knee deep in the ever-so inspiring "Envision San Jose."

The CityNature research group is exploring why access to nature (defined as tree-lined boulevards, urban parks, backyard gardens) doesn’t scale up with population. We took a variety of approaches, and I found myself on the Semantics Team. The Semantics Team sought to understand how cities talk about nature – Is nature aesthetic? Economic? Recreational? Or is it something else altogether? In doing so, my other undergraduate teammate and I built a plain text corpus out of the 37 comprehensive plans we were able to get our hands on. This turned out to be no small task. I knew from the beginning that I wanted to talk about the social causes and effects of the physical urban design of city nature, and I realized I could use this painstakingly pieced-together corpus just for that purpose.

My research has focused on what city comprehensive plans refer to as “social capital.” Social capital is a collective term for “social networks and the associated norms of reciprocity” – it is, essentially, the power of human relationships rather than the power of an individual or even a collective group (Putnam, 2000). Social capital lies between the influence and capital of a single person and the power in a great number of people – social capital refers to neither Martin Luther King, Jr. nor the sum of the March on Washington, but instead to the congregation that attended in full force because of the strong ties between its members. I know, I know, it’s a mouthful.

I’m working with the Digital Humanities specialists at Stanford — Karl Grossner and Elijah Meeks — to use topic modeling, which is an approach to natural language processing, to find out how and under what circumstances cities talk about social capital. Because social capital isn’t easily defined, the prep work for the data analysis has been the most consuming. I’m using three methods of looking at the corpus. First, I’ll be running generic topic modeling to see which “topics” are related to social capital. Then, each city will have a signature – each city talks about social capital x amount, and in y and z locations in their plan. The second and third methods involve comparing the corpus to the “quintessential social capital reader text” that I’ll be putting together from some choice sources.

So I’m happily spending my last few weeks here with the Bill Lane Center getting results. Real results! And after all, what’s more fun than that?

Read more at the Out West Blog for Summer Interns »