Data. Race and ethnicity: U.S. Census Bureau, American Community Survey 2020–2024 5-year estimates, table B03002 (Hispanic or Latino origin by race), census tracts in Los Angeles and Orange counties. White, Black and Asian mean non-Hispanic. College degree: share of adults 25+ with a bachelor's degree or higher (B15003). Married: share of residents 15+ who are currently married, including separated (B12001). Median household income: B19013, in 2024 dollars; the Census caps this figure at $250,001. Boundaries: Census TIGER/Line 2024 tracts. Highways: TIGER/Line 2024 primary roads (limited-access freeways and tollways, feature class S1100), clipped at the Los Angeles and Orange County lines and simplified for display; small labels give the route numbers. The dashed line is Mulholland Drive between the 405 and the 101 (TIGER/Line 2024 LA County roads), marking the ridge that separates the San Fernando Valley from the LA basin. Stores: Trader Joe's store locator (locations.traderjoes.com), checked 30 Oct 2026, 90 stores in the two counties (the Westlake Village store sits just over the line in Ventura County and is left out). Sprouts (60 stores) and Aldi (43 stores): StoreLocators.com chain lists for California, updated 29 Oct 2026, LA and Orange County entries. Erewhon (13 stores): StoreLocators.com list; the Thousand Oaks store is in Ventura County and the Pacific Palisades store is closed after the 2025 fire, so neither is shown. Smart & Final (98 stores): the company's official store directory, smartandfinal.com/store-locations, read 30 Oct 2026; LA and Orange County stores only. Ralphs (130 stores): Ralphs' own store locator (ralphs.com), with the company's store coordinates, read 30 Oct 2026. Vons (77 stores): Vons' official store pages (local.vons.com), with the company's store coordinates, read 30 Oct 2026. Median age: ACS 2020–2024 table B01002. Median home value and homeownership: ACS 2020–2024 tables B25077 (median value of owner-occupied homes; top-coded at $2,000,001) and B25003 (tenure: share of households that own their home). Food insecurity: Food Base LA (USC Spatial Sciences Institute, USC Dornsife Public Exchange and the LA County Office of Food Systems), 2023 tract layer: the share of households estimated to be food insecure in the past 12 months, from Feeding America's Map the Meal Gap model. Covers Los Angeles County only, so Orange County tracts show no data on that layer.

Method: placing stores. Trader Joe's, Sprouts, Aldi, Erewhon and Smart & Final addresses were placed with the Census TIGER 2024 address-range files: each address is interpolated along its street segment, so a pin can be off by up to about a block. A few stores with no house number or no matching range were placed at the named intersection or at the middle of that street within their ZIP code. Ralphs and Vons pins use the coordinates published on each company's own store pages. Each store is assigned to the census tract its pin falls inside; a few stores sit in tracts with no residents (commercial or industrial land) and drop out of tract-level comparisons.

Method: shading. Each tract is coloured by one ACS 2020–2024 value. Race layers show the share of residents in the chosen group in five 20-point bands. Every other layer uses four tiers chosen to split the region into meaningful, roughly comparable groups (for example income under $70k, $70–100k, $100–150k, $150k+). Tracts with no residents, or a suppressed estimate, are grey. Tract outlines are simplified for display.

Method: "Stores vs. residents" panel. For race, each tract is labelled by its largest group; the bars compare the share of a chain's stores in each kind of tract with the share of the two counties' residents living in that kind of tract. For the other layers the same comparison is made across the four tiers. If stores were spread in proportion to people, each chain's bars would match the grey residents bar.

Method: "Lives within 1–3 miles" panel. Each tract is reduced to its centre point. The straight-line distance from that point to the nearest store of each selected chain is measured; if it is within the chosen radius, every resident of the tract counts as "near." For race, the result is the share of each group's residents (LA and Orange counties) who are near a store. For the other groupings, tracts are grouped into the map's tiers and the bar is the share of residents of those tracts who are near a store. "Group by" sets the outer rows and "Then by" the bars inside each row; ⇄ swaps them without changing any number. One of the two is always the grocery chain, because the measure is distance to a specific chain's stores. Distances ignore roads, freeways and transit, and chains with more stores naturally score higher, so compare groups within a chain more than chains with each other.

Method: grocer comparison tab. Choose any two chains to compare head to head on 14 site-selection metrics. Most come from ACS 2020–2024 tract tables: population density (B01003 population over TIGER land area), population within 2 miles (all tracts whose centres are within 2 miles), median household income (B19013), households earning $150k+ (B19001), median home value (B25077), bachelor's degree (B15003), median age (B01002), ages 25–44 (B01001), married (B12001), one-person households (B11016), homeownership (B25003) and households with no vehicle (B08201). Rival grocers within 2 miles counts, for each chain, the stores of the other six chains on this page within 2 straight-line miles of each tract centre; the region baseline is computed the same way for each chain. Reported thefts within 2 miles are LAPD 2023 larceny and vehicle-theft reports (City of Los Angeles open data) per 1,000 residents across all tracts within 2 miles, shown only where at least 80% of those residents live in the City of LA; using the 2-mile area rather than the store's own tract keeps thefts at the store itself from dominating. For each metric, every populated tract in the two counties is compared with the tracts that hold each chain's stores (a tract with two stores counts twice; stores in tracts with no residents are left out). "Typical" is the median. The difference is (average of the chain's store tracts − average of all tracts) ÷ standard deviation across all tracts; density, store counts and thefts are heavily skewed, so their differences are computed on a log scale. Each chain's three largest differences are highlighted. Chains with few stores (Erewhon has 13) move a lot if one store opens or closes. Race is shown separately in the note at the bottom of the tab.

Method: average store tab. Site and building traits for all 511 stores were read from OpenStreetMap on 31 Oct 2026 (map data within about 200 m of each store): the store's mapped building and its outline area, other shops inside the same building, retail or mall land around it, parking mapped within 60 m, and the OpenStreetMap class of the street named in the store's address. Building outlines approximate store size (they include back-of-house and, for some stores, attached tenants). About a third of stores have no mapped building, so shares are of the stores where the trait is mapped. Census rows reuse the tract data on this page. The "most average" store is the one whose tract is closest to the chain's median profile. The four hybrid addresses are explained on the tab.

Limits. A store's shoppers come from well beyond its own tract, so the 1–3 mile figures are the fairer measure of access. The numbers describe where stores are, not why: income, education, home values and race overlap heavily in Los Angeles, and nothing on this page separates them. The companion spreadsheet compares 2-mile trade areas and estimates weights for Trader Joe's and Erewhon without race or food insecurity. Census figures are 2020–2024 averages, so recent changes (such as the January 2025 fires) are not reflected.