UN Data Commons × NYC

Briefing hub for the UN System Data Commons hackathon project

View the Project on GitHub sarapis/undatacommons-nyc

Crosswalk status — 2026-09-14

Both sides of every mapping, verified. The grade in each entry is a human judgment recorded in probe/crosswalk.json; this page verifies that the mapping still resolves and that the units agree.

Tier 1 = NYC vs national city aggregates · Tier 2 = urban aggregates · Tier 3 = national totals. The tier says what the comparison is worth.

Pair SDG Tier Grade Chartable Overlap Blockers
Fine particulate matter (PM2.5), annual mean 11.6.2 1 PROXY yes 2010–2019
Intentional homicide 16.1.1 3 PROXY yes 2006–2023
Proportion of municipal waste recycled 11.6.1 3 PROXY yes 2000–2018
Road traffic deaths 3.6.1 3 BLOCKED NO 2021–2021 grade BLOCKED: human judgment says these must not share an axis
Child mortality (deaths) 3.2.1 3 CONTEXT NO 2021–2023 grade CONTEXT: human judgment says these must not share an axis
Maternal mortality 3.1.1 3 BLOCKED NO 2016–2021 grade BLOCKED: human judgment says these must not share an axis
Deaths attributable to ambient air pollution 3.9.1 3 CONTEXT NO 2012–2017 grade CONTEXT: human judgment says these must not share an axis
Poverty 1.2.1 3 CONTEXT NO no NYC counterpart identified
Inadequate housing 11.1.1 2 CONTEXT NO 2010–2022 UNIT MISMATCH: UN Percent vs NYC count; grade CONTEXT: human judgment says these must not share an axis
Built-up area per capita (cities) 11.7.1 1 NO-NYC-SOURCE NO no NYC counterpart identified
Safely managed drinking water 6.1.1 3 NO-SIGNAL NO no NYC counterpart identified

Detail

Fine particulate matter (PM2.5), annual mean (SDG 11.6.2) — PROXY, Tier 1

Units agree (ug/m3) and both are annual means, but the methods differ: NYC’s figure comes from NYCCAS ground monitors, the UN’s is a population-weighted modelled estimate. Comparable in level, not interchangeable. Chart together with the method difference stated.

Intentional homicide (SDG 16.1.1) — PROXY, Tier 3

NYPD ‘murder & non-negligent manslaughter’ tracks the UNODC intentional-homicide definition closely. BLOCKER: NYC side is a count, UN side is a rate per 100,000 – needs an annual NYC population denominator before the two can share an axis. Pre-2006 rows in this dataset carry junk incident dates (17 murders in 1990) and are excluded. Denominator: Census ACS 1-year total population. There is no ACS 1-year release for 2020, so 2020 carries no rate – and 2020 is the year NYC homicides jumped. That gap is left empty rather than interpolated.

Proportion of municipal waste recycled (SDG 11.6.1) — PROXY, Tier 3

DSNY tonnage covers DSNY-collected residential waste only – it excludes commercial waste, which the UN municipal-waste definition includes. Derived diversion rate, not a published one; the published Recycling Diversion and Capture Rates dataset has not updated since 2020. The month field is text (‘2026 / 08’) so the year comes from substring, not date_extract_y. Rows before 1993 predate curbside recycling (no recycling tonnage, 1990 partial) and are excluded.

Road traffic deaths (SDG 3.6.1) — BLOCKED, Tier 3

NYC has an excellent series (Vision Zero collision data, daily updates). The UN side has ONE observation, 2021, a periodic WHO modelled estimate. Kept in the crosswalk deliberately: this is what a blocked pair looks like, and the blocker is the UN side, not ours.

Child mortality (deaths) (SDG 3.2.1) — CONTEXT, Tier 3

AGE BANDS DO NOT MATCH. The UN counts deaths under five; NYC’s infant mortality dataset counts deaths under one. Both are raw counts so the units ‘agree’, which is exactly the trap – a chart would look correct and compare different populations. Show side by side, never on one axis, until an under-five NYC series is sourced.

Maternal mortality (SDG 3.1.1) — BLOCKED, Tier 3

Two problems. NYC publishes pregnancy-ASSOCIATED deaths (any cause within a year of pregnancy), a deliberately broader definition than the UN’s maternal mortality. And the UN rate is per 100,000 LIVE BIRTHS, not per population – the denominator lives in a different NYC dataset (fcau-jc6k number_of_live_births). Needs a cross-dataset join before it can be charted.

Deaths attributable to ambient air pollution (SDG 3.9.1) — CONTEXT, Tier 3

UN reports a national death COUNT attributable to ambient air pollution; NYC reports an age-adjusted RATE for ages 30+ attributable to PM2.5 specifically. Different measure, different age base, different pollutant scope. Informative side by side, not comparable on one axis.

Poverty (SDG 1.2.1) — CONTEXT, Tier 3

NOT THE SAME CONCEPT. The UN series is the share below the INTERNATIONAL extreme-poverty line (~$2.15/day PPP), which is near zero for any US city and carries no signal. NYC’s own poverty measure is a cost-of-living-adjusted threshold an order of magnitude higher. Also the NYCgov measure is published as a separate dataset per year, so there is no continuous series to query. Included as the clearest example of two numbers that share a name and measure nothing alike.

Inadequate housing (SDG 11.1.1) — CONTEXT, Tier 2

UN measures the SHARE OF URBAN POPULATION living in slums. NYC’s nearest analogue is a COUNT of class C (immediately hazardous) housing maintenance violations – an enforcement output, not a population share, and sensitive to inspection rates rather than to conditions alone. Directionally interesting, structurally different. Tier 2: the UN side is urban, which bundles cities with suburbs.

Built-up area per capita (cities) (SDG 11.7.1) — NO-NYC-SOURCE, Tier 1

Tier 1 UN data exists with a full 2000-2025 series and city-level slices across countries, but no NYC Open Data equivalent has been identified. Candidates to investigate: PLUTO lot area aggregates, Parks Properties. Recorded so the gap is visible rather than forgotten.

Safely managed drinking water (SDG 6.1.1) — NO-SIGNAL, Tier 3

The UN series sits at or near 100% for the US across the whole window, as it would for NYC. A chart of two flat lines at 100 tells a policy analyst nothing. Recorded as deliberately excluded rather than overlooked – a ceiling effect is a reason not to build a view, and that judgment should be written down.

Summary

11 pairs · 3 chartable end to end.

A blocked or context-only pair is not a failure. Recording why two numbers cannot share an axis is the product; a crosswalk showing only the easy pairs would be the thing we are building against.