Henry Gomory

Facebook Twitter Google Print Friendly and PDF

Tell us a little about yourself.

I grew up in Princeton, New Jersey and recently graduated from Harvard College, studying sociology. I care deeply about issues of social inequality and I am drawn to social science research for its ability to explain social problems and reveal societal injustice.

Can you describe your honors thesis?

Proportion of renters in each census tract spending more than 50% of their income on rent

For my honors thesis, I used Boston Neighborhood Survey and 311 data to show that neighborhoods in which residents spend a high percentage of their income on rent have smaller social networks, less trust between neighbors, and less custodial behavior. I also used 911 records, property assessment data, and a number of census indicators to test whether this relationship was the spurious result of other aspects of rent-burdened neighborhoods.

How did you get involved with BARI?

I took several years off in college and worked as a research assistant at Columbia University. When I returned to Harvard, professors at Columbia put me in contact with BARI Co-Director Robert Sampson, who hired me to work at BARI as a research assistant.

Give us a brief overview of the projects you’ve worked on for BARI.

My main role at BARI has been preparing large administrative datasets for research. I have worked with 311 Constituent Relationship Management data, 911 call records, building permits, business licenses, property assessments, zoning clearances, the Boston Neighborhood Survey, and the geographical infrastructure.

Tell us about the Geographical Infrastructure. How did it come to be? What does it measure?

The geographical infrastructure is a group of nested datasets that organizes the physical landscape of Boston. Beginning with the smallest ownable pieces of property, it moves up to parcels of land, then road segments, census geographies, and neighborhoods. Its purpose is to be a definitive geography of Boston, so that other datasets can be placed within that infrastructure. One key function is connecting city records to census geographies.

Is there something that makes this release different than previous versions?

We are constantly reevaluating the geographical infrastructure to ensure that it reflects Boston’s geography as accurately as possible. In this release, we improved our method of geocoding a property to a street segment by using latitude and longitude data to choose between possible segments. Previously we had relied on the street numbers provided by the US census, which are largely accurate, but are nevertheless only estimates. Using geographic distance allowed us to choose the closest street segment that matched on name and avoid faulty matches based on estimated numbers. Also, the 2017 version is the first to include a mappable shapefile for land parcels, which enables the visualization of address-level characteristics with actual building footprints.

Potential for use in research, policy, practice?

The geographical infrastructure allows us to place our other datasets within a larger geographical context. This allows us to answer questions like how physical construction varies across neighborhoods, where businesses are starting, or what properties are hotspots for 311 requests.

What’s next for you?

This fall I will begin a PhD program in Sociology at Princeton, where I plan to focus on housing inequality and affordability.


If you have questions about Henry’s work or would like to get in touch with him, email us at

To read more about the release of the Geographical Infrastructure (v. 2017), click here.

Published On: August 29, 2017