Map Alert!

Some more interesting patterns of social connections across U.S. zip codes (Thanks to @michaelcbailey, @pfrrll and Theresa Kuchler for exploring these with me over the years).

I learn something new every time I play with these data.

(1/N)

The Social Connectedness Index (SCI) captures the likelihood that individuals in two different regions are friends on Facebook.

Our social networks have tremendous influence on our lives and capture both lived realities and a surprising degree of history.

(2/N)

The SCI highlights the persistence of social ties over long distances, enables analysis of the radically differing aggregate social networks of even nearby locations, and can light on the impacts of segregation (as captured by social networks) within cities.

(3/N)

One of the largest mass movements of people in the US was the Great Migration, when large numbers of Black Americans from the South moved to cities in the North, Midwest, and West beginning in the early 1900s.

(4/N)

This migration reshaped the demographics of many cities, and social networks today still reflect these migration patterns.

Looking at majority Black zip codes in Chicago (60653), Cleveland (44104), and New York (10030), we see areas with strong connectivity in the South.

(5/N)

Considering again these majority-Black zip codes, notice that Chicago’s southern connections are strongest near Louisiana and Mississippi, New York’s in South Carolina, and Cleveland’s somewhat in-between these two distributions.

(6/N)

These distributions mirror Great Migration patterns, where migrants from the South would often travel north to the nearest large city by rail.

https://en.wikipedia.org/wiki/Great_Migration_(African_American)

(7/N)

Nearby majority white zip codes, however, do not show these social ties to areas of the South.

What the SCI is picking up is the persistence of unique linkages of Black Americans to these areas.

(8/N)

Other notable migration patterns also appear in the SCI data. The Bakersfield area in Kern County, California, was a major destination for mostly white migrants fleeing the Dust Bowl.

After subsequent demographic change, many areas of the city are majority Hispanic.

(9/N)

A majority white zip code near Bakersfield (Oildale, 93308) shows strong ties to the Oklahoma/Arkansas/Texas area, while a majority Hispanic zip code (93307, in Bakersfield proper) in the same area does not.

(10/N)

The Great Migration also reached Los Angeles, and social linkages are still evident, mostly in the western-most areas of the South, reinforcing the earlier pattern.

(11/N)

Looking within urban areas, we can also pick out some unique patterns of social networks.

For the within-CBSA/county/city figures to follow, the scale will reflect the degree of connection (90th percentile, 80th percentile, etc) to other zip codes in the area shown.

(12/N)

For instance, Beverley Hills, and its famous 90210 zip code, shows strong connections to wealthy, majority-white resort areas further south (Huntington Beach and Laguna Beach) but weak connections to nearby Hispanic- and Black-majority zip codes.

(13/N)

These majority-Hispanic zip codes (e.g., 90003), show the reverse pattern.

Connections to the west (including affluent neighborhoods such as Beverly Hills, Santa Monica, and Malibu) are weak.

Connections to distant Hispanic-majority zip codes to the north are strong.

(14/N)

Pittsburgh: of the zip codes with the strongest connections to the Hill District (a majority Black neighborhood near the city center), NONE are directly adjacent

Its strongest connections are to majority-Black communities separated by rivers or outside of the city proper

(15/N)

Zip codes close by with weak connections to the Hill District include those home to major universities (these are the zip codes with lower than expected connectivity directly to its east).

(16/N)

Segregated social networks are also evident in New York.

A zip code in Central Harlem has strong connections to the Bronx and to majority-Black areas in Brooklyn and Queens, and weak connections to much of the majority-white areas in the UES and South Brooklyn.

(17/N)

Looking at the network of a zip code in the majority-white Upper East Side, we see weak connections to the Bronx and strong connections to majority-white zip codes nearby in Brooklyn and Queens.

(18/N)

Overall, I am always amazed how much U.S. and world history we can learn by looking at these maps, and trying to understand what is going on.

The Social Connectedness Index is an endless source of interesting maps.

(19/N)

We explore a number of additional patterns within NYC in this paper here:

http://pages.stern.nyu.edu/~jstroebe/PDF/BFKS_SocialConnectednessUrbanAreas_JUE.pdf

(20/N)

I have summarized some previous findings on social connectedness in Europe here

(21/N) https://twitter.com/stroebel_econ/status/1284540562746548225?s=20

And findings on social connectedness within Germany here

(22/N) https://twitter.com/stroebel_econ/status/1288127309841485825?s=20

And @rrichmond wrote a great thread on our work using the Social Connectedness Index across countries

(23/N) https://twitter.com/rrichmond/status/1285962440350588932?s=20

The data is accessible to other researchers (email [email protected]), and people have used it to study many interesting social phenomena in economics and finance!

(24/N) https://twitter.com/stroebel_econ/status/1247537135487299590?s=20