2020

The Border Between Red and Blue America

Category: Best data-driven reporting (small and large newsrooms)

Country/area: United States

Organisation: The New York Times

Organisation size: Big

Publication date: 25/10/2019

Credit: Robert Gebeloff, Sabrina Tavernise, Christopher Lee, Lauren Leatherby, Jugal Patel

Project description:

The American political landscape is polarized, and most political reporters identify the suburbs as an important electoral battleground. But a New York Times analysis found that the border between conservative and liberal America is more pronounced than the pundits realized. In fact, draw a line through the suburbs — a line dividing the older suburbs built during the post-World War II era and the newer suburbs that have sprung up recently, and you’ll find two distinct demographic and political worlds.

Impact reached:

These stories for the first time quantified the extent to which the U.S. has changed politically and demographically. For decades, the assumption has been that the great socieo-economic dividing line in the U.S. is between cities and suburbs. Many have sensed that the suburbs are becoming more like cities. However, these stories for the first time documented this trend with data and established that, in fact, the “suburbs” has truly become two categories. The newer, “outer ring” suburbs are totally different in almost every way from the older, “inner ring” suburbs.

Techniques/technologies used:

The U.S. Census Bureau has no official definition of a “suburb”, and so academic researchers are left to devise their own methods. After reviewing these efforts, we came up with our own version, building on previous work. We started with neighborhood-level population density — the U.S. is divided into more than 74,000 Census “tracts.”  But the problem with this approach is that there are many industrial neighborhoods in urban places with low population density — few live there, but they are clearly urban. So we took a second approach, using satellite imagery from the Multi-Resolution Land Use Consortium. The satellite imagery presented the U.S. as a set of several million pixels, with each pixel color-coded to represent land density. We wrote a script that joined the pixels to Census tracts, which then enabled us to create a development density index for each tract. So for each tract, we had population density and development density, which we were able to combine into an overall index of urbanness, on a scale of 1 (most rural) to 10 (most urban).  Our original plan was to call any neighborhood with a 1 or 2 rural, 9 or 10 urban, and everything in between suburban.

But those plans changed when we started working with data. We then created a database of demographic information at the tract level spanning 40 years. We also brought in neighborhood-level election results from the 2016 presidential race. When we grouped these data by our density index, the pattern was clear: Rural was definitely distinct, as was Urban. But our Suburban category had to be split as well – 2 through 5 were totally different than 6 through 8. 

We then set about confirming these findings through traditional reporting – on the street and through experts.

What was the hardest part of this project?

The project presented some major technological hurdles — namely, figuring out how to convert a gigantic, high-density pixel map into something that would spatial-join with standard GIS shape files.

But the bigger challenge is then taking complicated findings and turning them into journalism that people will actually want to read. The first story was a partnership with colleague Sabrina Tavernise, who spent weeks on the ground in Kent County, Michigan, getting people in various neighborhoods to open up about their political beliefs. In this day and age, such conversations are not easy in the U.S., but Sabrina perservered and got perspectives from across the political spectrum.  The photos of Christopher Lee were especially effective at bringing the findings to live.

After a surprising local election in the State of Virginia, Sabrina once again set out to tell the story about the changing suburbs, this time on a tight deadline,  while I conducted additionanl layers of analysis specific to this state — this is the third story in the package.

But the bottom line is, our analysis was complicated and difficult to “storify” — but I think this represents a perfect marriage between advanced data techniques and traditional storytelling.

What can others learn from this project?

The two main takeaways: Taking on the common knowledge of pundits is always a newsworthy exercise. In this case, our big get was showing how commentators talking about the “suburbs” as a single entity were in fact over-simplifying reality, and that reporters need to keep the inner ring – outer ring divide in mind at all times.

The second is the importance of narrative journalism. One of the stories in the series is my own data column about the findings. But this effort would not have been nearly as effective without the other two pieces, which allowed us to tell the story from a wide variety of perspectives.

Project links:

www.nytimes.com/2019/10/25/us/democrats-republicans-suburbs.html

www.nytimes.com/2019/10/29/upshot/suburbs-demographics-red-blue.html

www.nytimes.com/2019/11/09/us/virginia-elections-democrats-republicans.html