Seven Chicagos, Defined

Anyone who knows anything about Chicago has heard some version of the “tale of two cities” metaphor. Some explain it rather curtly (“there’s an invisble border at Roosevelt Road”), others have described it in much more detail. But the core idea is always the same: there are two very distinct parts of Chicago; one is extremely rich, White and highly-educated, resembling the Upper East Side/Upper West Side of Manhattan, northern San Francisco or northern Atlanta. The other is extremely poor and mostly Black, like Detroit or Cleveland.

My baseline take on “the two Chicagos” is:

1. It’s true: both of these two areas exist, and they’re not that different from how they’re described. And while most other cities in the eastern United States (and almost all large US metros) have the same issue to some extent, Chicago is not unfairly singled out, it’s the most segregated large city in America.
2. However, these “two Chicagos” do not encompass anywhere near the entirety of the City; they take up only a fraction of the 77 neighborhoods and 2.7 million residents that comprise the capital of the Midwest. A plurality of Chicagoans live in neither of the two types of neighborhoods that people typically think of when they use this metaphor.

Establishing a more nuanced view of Chicago’s demographic divides

My view is that in order to capture the key fault lines of Chicago’s 77 neighborhoods, you need at least 6 or 7 buckets, not 2. So to get an idea of what that would look like, I assembled a summary of the demographic breakdown of each neighborhood using data from the CMAP data hub. I looked at four different demographic axes: race/ethnicity, age, income, and education. The segmentations used are as follows:

Race: Non-Hispanic White, Non-Hispanic Black, Hispanic, Asian

Age: <20, 20–34, 35–49, 50–64, 65–74, 75–84, 85+

Household Income: <25k, 25–50k, 50–75k, 75–100k, 100–150k, >150k

Educational attainment: Less than High School, High School diploma, Some college, Associate’s degree, Bachelor’s degree, Graduate or Professional degree

Next, I fed these segmentations into a k-means clustering algorithm to categorize the neighborhoods into seven different groups.

Note that there are no geospatial variables, so the neighborhood groupings do not have to be geographically contiguous, and when they are, it’s purely a function of demographics. Density and transit aren’t variables either, but I’ll still reference patterns that emerge across those axes, since density and transit no doubt have an impact on demographic trends, and vice-versa.

Some caveats up-front

This is not an attempt to create any sort of definitive neighborhood grouping; For one, k-means is not a deterministic algorithm — you can run it multiple times with the same data and get different results. In addition, the choice of 7 clusters is relatively arbitrary — you could make an argument that 6 or 8 or even 10 clusters would be more appropriate for various reasons. And of course, there are many other variables you could include to get a more well-rounded picture that goes beyond basic demographics.

The main goal here is to show that you can split Chicago’s neighborhoods up into 7 groups and still be able to tell a coherent story about each of them —there’s enough variation that you won’t get multiple groups that are extremely similar.

The Map

The result I got running k-means clustering with seven clusters looks like this:

The seven Chicagos

Anyone familiar with Chicago will immediately recognize some patterns: the red neighborhoods are the mostly-White, wealthy and highly-educated; the yellow neighborhoods are predominately Black; and the green neighborhoods are all plurality-Hispanic. However, there are other large groupings that don’t cleanly fit into any of these archetypes: a cluster of orange on the far-northern lakefront, large areas of purple on the northwest and southwest sides, and a chunk of blue on the far northwest edge of town. In addition, there are islands of pink, green, orange, blue, and purple dotted around the edges of the south side.

Let’s breakdown each of these “seven Chicagos” one-by-one:

Region 1: White Suburban Chicago (Blue on Map)

Neighborhoods: Dunning, Edison Park, Forest Glen, Jefferson Park, Lincoln Square, Mount Grenwood, North Park, Norwood Park, O‘Hare
Population: 232,000
Density: 6,100 residents / square mile
Population change since 2010: + 8,900
Foreign-born population: 24%

These neighborhoods are all majority-White, relatively sparse, and generally have older age distributions. Eight of the nine neighborhoods are geographically contiguous on the northwest corner of town; the exception is Mount Greenwood, which is located on the southwest edge of town and was the only neighborhood that Donald Trump won in 2016, mostly because lots of cops live there.

None of these neighborhoods are poor, but median income varies quite a bit; O’Hare and North Park are both near the citywide median of ~50k (North Park is the most “economically integrated” neighorhood in Chicago), while Edison Park and Forest Glen have among the highest median incomes in the city at 101k and 111k, respectively.

One neighborhood that stands out here is Lincoln Square; it’s the youngest of the group and has the highest population density. I imagine it was close to being grouped with neighboring North Center and other younger neighborhoods.

Region 2: Black Chicago (Yellow on Map)

Neighborhoods: Auburn Gresham, Austin, Avalon Park, Burnside, Calumet Heights, Chatham, East Garfield Park, Englewood, Fuller Park, Grand Boulevard, Greater Grand Crossing, North Lawndale, Oakland, Pullman, Riverdale, Roseland, South Chicago, South Shore, Washington Heights, Washington Park, West Englewood, West Garfield Park, West Pullman, Woodlawn
Population: 606,000
Density: 9,700 residents / square mile
Population change since 2010: — 36,400
Foreign-born population: 3.5%

Region 2 is comprised of a whopping twenty-four neighborhoods, all on the south and far-west sides, encompassing almost all of Chicago’s majority-Black neighborhoods. Together, they are home to over half a million Black residents: that’s more than all of Montana, Wyoming, Vermont, North Dakota, Idaho, South Dakota, New Hampshire, Maine, Hawaii, Alaska, Utah, New Mexico, West Virginia, Oregon and Rhode Island combined (think about what that means in terms of representation in the US Senate).

When people talk about the “two Chicagos”, most of the neighborhoods they picture as representing “poor Chicago” are in this region. Thirteen of these neighborhoods have a median income below 30k, and most of the neighborhoods with the highest violent crime rates are on this list. However, a good chunk of this list is near the city-wide median income; most of them are near the city’s southwest border, including Washington Heights, Pullman, Roseland and West Pullman.

I was surprised that the wealthier Black neighborhoods on the southwest and southeast sides were grouped together with the far-west side and central-south side neighborhoods, and they might not have been if I had chosen to make 8 or more clusters. The key factor causing the clustering algorithm to keep this huge menu of neighborhoods together was likely the sheer level of segregation: 90% of the combined population of these neighborhoods is Black. The other 3 demographic axes vary a decent amount, but the racial element does not.

Region 3: Yuppie Chicago (Red on Map)

Neighborhoods: North Center, Lake View, Lincoln Park, Near North Side, The Loop, West Town
Population: 417,000
Density: 23,900 residents / square mile
Population change since 2010: + 34,500
Foreign-born population: 14%

This is the booming bubble. Like the neighborhoods in region 1, these six neighborhoods are all majority-White. However, this hyper-dense, 17.5 square mile area differentiates itself from region 1 with a younger, more educated and higher-income population. A whopping 78% of adults over 25 have a Bachelor’s degree, 78% of residents are under the age of 50, and median incomes range from 89k (Lake View) to 116k (North Center).

In the “tale of two cities” paradigm, these are the neighborhoods people picture as representing the rich half of the city. And indeed, they fit the stereotype extremely well. But at about 417,000 residents, these neighborhoods make up only 15% of Chicago’s population, even after adding over 34,000 residents in the past decade.

Region 4: Somewhat Integrated Chicago (Orange on Map)

Neighborhoods: Rogers Park, Edgewater, Uptown, Near West Side, Near South Side, Hyde Park, Beverly
Population: 305,000
Density: 16,800 residents / square mile
Population change since 2010: + 15,500
Foreign-born population: 22%

The first three regions all had pretty clear-cut patterns, and were mostly geographically contiguous. This seven-pack of neighborhoods, however, is all over the place geographically, and is much more diverse. In fact, with a racial distribution of 49% White, 23% Black, 13% Hispanic and 12% Asian, this region features a majority of the ten most racially integrated neighborhoods in Chicago.

Some of this on-paper integration comes with big caveats. Near West Side has the 4th largest land area of any neighborhood at 5.7 square miles, and it has some obvious bifurcations: the east edge is more-or-less a continuation of the Loop from region 3 (hence the “west loop” nickname), while everything west of the United Center more closely resembles neighboring East Garfield Park from region 2, which is 87% Black. Near South Side has the same general dynamic, albeit at at smaller scale. And anyone who has ever been associated with Hyde Park and/or the University of Chicago knows that the relationship between the majority White/Asian University and the area’s Black residents is complicated.

However, I do think there is a lot of good stuff going on in these neighborhoods. The northern lakefront trio of Uptown, Edgewater and Rogers Park are reasonably diverse across all four demographic axes, while also having high population density and transit usage (Rogers Park is the most racially integrated neighborhood in Chicago, Edgewater is 5th, Uptown 8th). And the University of Illinois at Chicago, located in Near West Side, is fairly diverse.

The southwest edge-neighborhood of Beverly is a sparse, fairly-wealthy enclave and would likely be grouped with region 1 if it wasn’t for that fact that it has a 35% Black population share, the highest of any majority-White neighborhood.

Region 5: Hispanic Chicago (Green on Map)

Neighborhoods: Archer Heights, Belmont Cragin, Brighton Park, East Side, Gage Park, Hermosa, Humboldt Park, Lower West Side, New City, South Lawndale, West Elsdon, West Lawn
Population: 480,000
Density: 13,800 residents / square mile
Population change since 2010: — 12,000
Foreign-born population: 34%

These twelve neighborhoods on the northwest, southwest and southeast sides make up most of Chicago’s plurality-Hispanic neighborhoods, and together are home to about 375,000 of the nearly 800,000 Hispanic Chicagoans. With a 34% foreign-born population share, these neighborhoods are a big reason why Chicago has the tenth-most foreign-born residents of any city in the world.

Median incomes range from 33k (New City) to 55k (West Elsdon), making all of these neighborhoods economically distinct from both the yuppie neighborhooods of region 3 and the poorer half of the region 2 neighborhoods. Likewise, from a density and transit perspective, these neighborhoods are in an intermediate zone between the highly-sparse regions 1 and 2 and the highly-urbanized regions 3 and 4. This region is also fairly young: it has the highest under-20 population share at 30%.

Region 6: Southside grab-bag (Pink on Map)

Neighborhoods: Morgan Park, Ashburn, Chicago Lawn, South Deering, Kenwood, Douglas
Population: 171,000
Density: 6,800 residents / square mile
Population change since 2010: + 600
Foreign-born population: 16%

At first glance, I was tempted to name this region “Black, middle-class Chicago”. And some of these neighborhoods do fit the bill: Morgan Park, Ashburn and Kenwood have median incomes of 68k, 61k, and 49k respectively and have a combined 56% Black population share.

However, this pattern does not entirely fit: a true “Black middle-class” grouping would need to include most-or-all of the southwest cluster of Washington Heights, Pullman, West Pullman and Roseland, as well as the southeast duo of Calumet Heights and Avalon Park. Those six neighborhoods, all part of region 2, have higher median incomes than Chicago Lawn, South Deering and Douglas, and have higher Black population shares.

I suspect the distinguishing feature that resulted in the algorithm clustering these neighborhoods together is that they are the only neighorhoods that are both plurality-Black and have a Black population share of less than 75%.

The source of integration varies — the lakefront neighborhoods of Douglas and Kenwood have significant White and Asian population shares; Meanwhile Ashburn, Chicago Lawn and South Deering have large Hispanic populations. Morgan Park is quite suburban and has a 28% White population share, the largest of any majority-Black neighborhood.

Region 7: California Chicago (Purple on Map)

Neighborhoods: Albany Park, Armour Square, Avondale, Bridgeport, Clearing, Garfield Ridge, Hegewisch, Irving Park, Logan Square, McKinley Park, Montclare, Portage Park, West Ridge
Population: 506,000
Density: 14,200 residents / square mile
Population change since 2010: + 11,000
Foreign-born population: 31%

In my opinion, this is the region that most contradicts the “tale of two cities” paradigm. Only one of these thirteen neighborhoods has a median income below 45k, and none are greater than 76k. Together, these neighborhoods have a population similar to Atlanta, Miami and Kansas City; they’re denser than Boston and DC; and they are fairly representative of Chicago as a whole across age, income and education. That’s a big, diverse, urban population, clustered mostly in 3 contiguous areas on the northwest and southwest sides.

The one area where this region fails to represent Chicago is the lack of Black residents. The racial demographics are 43% Hispanic, 36% White, 15% Asian, and 5% Black. That’s strikingly similar to the demographics of California. And not-coincidentally, this is also the one region that sees the same debates around development and gentrification that the Golden State is famous for, especially around the northwest side neighborhoods of Logan Square, Avondale and Irving Park, all of which border yuppie-dominated North Center and Lincoln Park.

Summary: where the two-Chicagos paradigm fails

The seven Chicagos demonstrate a number nuances that the “tale of two cities” paradigm fails to capture. The key highlights:

1. Hispanics make up almost a third of Chicago, and most plurality-Hispanic neighborhoods are clearly distinct from the both the ultra-wealthy neighborhoods and the ultra-poor neighborhoods.
2. There are large swaths of the city that demonstrate California-esque integration between the Hispanic, White and Asian populations, and those neighborhoods are economically representative of the city as a whole.
3. Integration between Black and White Chicagoans is pitiful, but there are exceptions, and it’s not just Hyde Park; Both the northern lakefront and the southwest side have multiple neighborhooods with high levels of integration.
4. Rich yuppie Chicago is very clearly a thing, but it makes up only 7% of the neighborhooods, 8% of the land area and 15% of the population of Chicago.
5. Over a million Chicagoans live in neighborhoods with a median income between 40k and 70k. The middle class is shrinking everywhere, including Chicago, but it very much still exists.

Bonus: Some data snapshots

Here are the 15 most racially integrated neighborhoods:

These are the most economically integrated

These neighborhoods have the lowest car ownwership rates. Notice there are only 2 rich neighborhoods on this list! Rich people in Lake View and Lincoln Park take transit to work, but they still own way more cars than poor people in Riverdale, Fuller Park and Washington Park.

Here are neighborhoods with the highest combined walk/bike/transit commute mode share. Not a lot of surprises here, honestly. But take a look at Edgewater, Uptown and Rogers Park: those are three diverse neighborhoods, all pulling greater than 50% non-car mode share.

And finally, here are the 7 cluster centers from the k-means algorithm. The cluster numbers on the left edge line-up with the region numbers used in this article.