Chicagoland's Activity Space Divide
A 2026 study in Demography found that neighbors of different incomes rarely share the same everyday spaces, even living blocks apart. This asks the same question at the town level across Chicagoland: when two municipalities border each other, do they actually share nearby cafes, libraries, and community spaces, or does each effectively have its own separate world?
Where the divide is sharpest
Each line connects two adjacent municipalities. Darker lines mean less overlap between what's actually reachable from each town, a proxy for shared activity space, not observed behavior (see the methodology note below for why that distinction matters).
What actually predicts the divide
Racial dissimilarity between adjacent towns, income gap, and the size mismatch between their catchments were tested together in a regression. Only two came out as real, independent predictors:
That's a real divergence from the original study, worth stating plainly rather than glossing over: the individual-level research centered income specifically. At the town level, using this structural proxy, race is what shows an independent effect and income doesn't. The overall model explains a modest but real share of the variation (R² = 0.042) — this is one real piece of a much bigger picture, not the whole explanation.
Tinley Park and Matteson are direct neighbors with a racial dissimilarity of 71 (one of the highest in the region) but an income gap of just $5,202 — about as close to income-equal as adjacent towns get here. Their third-place catchments share zero overlap. If income drove this, these two towns should look demographically alike. They don't.
Real neighbors, by name
Methodology & honest caveats
Geography: 319 Census-designated places across Cook, DuPage, Kane, Lake, McHenry, and Will counties. Adjacency determined by actual boundary proximity (within 50m, to catch places separated by a thin road), 609 adjacent pairs found.
Catchment: all OSM third places (cafes, libraries, community centers, social facilities) within 1.5 miles of each municipality's boundary. Overlap is measured as the Jaccard index between two adjacent towns' catchments (shared third places ÷ total distinct third places across both).
The core limitation, stated plainly: this measures opportunity for shared activity space, what's physically nearby both towns, not observed behavior, which is what the original individual-level mobility study actually measured. Two towns could share a large catchment on paper and still have residents who never actually cross into it. This is a structural proxy for the real question, not a replication of it.
Demographics: 2023 5-year American Community Survey estimates (median household income; White alone, Black alone, and Hispanic population shares) by place.
Data pulled August 2026.