UN Data Commons × NYC Open Data · Builders' Day, Google NY, 22 Sep 2026
New York published the world's first Voluntary Local Review in 2018. The UN now holds the authoritative global series. Joining them is the easy part. Knowing which joins are legitimate is the whole job — and most of them are not.
Every figure below carries its source, its vintage and its unit. Hover any chart to read the year.
6.60 µg/m³ in 2019, against 7.57 for the US city aggregate — and the gap has widened every year since 2016.
The UN publishes a degree-of-urbanisation slice, so this compares NYC to the city parts of the US rather than to the country. Only five indicators have it.
Only four countries' cities breathe cleaner air than New York's — Iceland, Guadeloupe, Finland and Estonia. This is the one comparison here that is like-for-like: Tier 1 means NYC is measured against other countries' city aggregates, not against whole nations.
Units agree (µg/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.
UN undata/sdg/EN_ATM_PM25.URBANIZATION--DOU_CITY · 10 obs 2010–2019 · unstats.un.org/sdgs/dataportal
NYC c3uy-2p5r Air Quality and Health Impacts · citywide annual mean · updated 2026-06-18
NYC ran above the national rate until 2013, then below it — and stayed below through the 2021 spike. Invisible from either dataset alone.
There is no ACS 1-year release for 2020, so there is no denominator and no rate — on the year NYC's count jumped to 456. We left it empty.
Against the United States, NYC looks like a success story. Against the world it sits in the bottom half — between Pakistan and Montenegro, with 84 countries reporting a lower rate. Same number, different comparator, opposite conclusion.
NYPD records murder & non-negligent manslaughter, which tracks the UNODC intentional-homicide definition closely. NYC publishes a count, not a rate — the denominator is Census ACS annual population. Pre-2006 rows in the NYPD file carry junk incident dates (17 murders in 1990) and are excluded.
UN undata/sdg/VC_IHR_PSRC · 21 obs 2000–2023 · per 100,000 population · unstats.un.org/sdgs/dataportal
NYC qgea-i56i NYPD Complaint Data Historic, offence code 101 · updated 2026-04-28
Denominator Census ACS 1-year B01003_001E · place 51000 / state 36
NYC recycles 17.2% against 23.6% for the US in 2018 — a six-point gap, and an easy headline.
The two numbers do not cover the same waste. The gap is real but overstated, and by an amount nobody has quantified.
Between Bahrain and Greece, and behind Japan and Germany. Only 63 countries report this indicator at all, so the field is thin — and the definitional caveat below applies to every one of these comparisons, not just the US one.
And one figure here is impossible. Malaysia reports 147.7% of its municipal waste recycled in 2019 — in the authoritative UN database. We left it in the data and clipped it off the axis rather than deleting it, because a silent drop is how a dataset launders its own errors. It is marked in red at the right edge.
DSNY tonnage covers residential collection only. The UN municipal-waste definition includes commercial waste. Same units, same period, same apparent shape — and the comparison is still not apples to apples. Every automated check passes here. Only a person reading both definitions catches it.
UN undata/sdg/EN_MWT_RCYR · 19 obs 2000–2018 · percent · unstats.un.org/sdgs/dataportal
NYC ebb7-mvp5 DSNY Monthly Tonnage Data · rate derived from tonnage · updated 2026-09-10
Note the published Recycling Diversion & Capture Rates dataset has not updated since 2020
We first marked road deaths blocked because the UN holds a single observation. That is true of a trend and irrelevant to a ranking — WHO publishes one global round, and 195 countries in one year is more than enough to place a city. The blocker was our framing.
NYC: 11 years of rates, every fatality geocoded, updated daily. The UN: 1 point, 2021. One line, one dot — on the same axis.
3.53 in 2013 down to 2.75 by 2018, then back up to 3.51 in 2021 and 3.16 in 2024. Visible only because NYC keeps the series the UN does not.
NYC's streets are as safe as Spain, the Netherlands and Germany, and roughly four times safer than the United States as a whole (14.2 per 100k). That finding is invisible while the comparator is the country NYC sits in — which is the argument for this card existing at all.
A rank is not a trend. NYC can be placed among 195 countries in 2021 and cannot be tracked against them in any other year. NYC counts deaths from collision records; WHO's figures are modelled estimates. And this is Tier 3 — a city set against whole nations, which is context rather than a peer comparison.
UN undata/sdg/SH_STA_TRAF · 1 obs 2021 · per 100,000 population · 195 countries reporting · unstats.un.org/sdgs/dataportal
NYC h9gi-nx95 Motor Vehicle Collisions — Crashes · persons killed by year · updated 2026-07-31
Denominator Census ACS 1-year B01003_001E · no 2020 release, so no 2020 rate
The United States does not report municipal waste collected to the UN — zero observations. Ninety other countries do. So this is the case where a city can be placed among the world precisely because its own country is absent from the table.
From 453.6 kg per person in 2005 to 387.6 in 2024 — about 15% less waste per New Yorker over two decades.
Not a gap in our pipeline. The UN holds 0 observations of this indicator for the United States, across every year.
Mid-table, between Hungary and Belarus. Read the caveat before believing it.
DSNY collects from residences and institutions only; commercial waste goes to private carters and never appears here. The UN definition includes commercial waste. So NYC's 397.9 kg is an undercount, its true figure is higher, and its true rank is worse than 41st — by an amount this page does not quantify.
Note the direction. The same residential-only gap makes NYC look worse on the recycling chart above and better here, because it sits in the numerator there and the numerator alone here. One definitional difference, two opposite distortions, on one page.
UN undata/sdg/EN_MWT_COLLV · tonnes · 90 countries in 2019 · zero US observations · unstats.un.org/sdgs/dataportal
NYC ebb7-mvp5 DSNY Monthly Tonnage · all collection streams summed · updated 2026-09-10
Derived per capita from Census ACS 1-year B01003_001E; country per capita from undata/unicef/DM_POP
Unit assumption DSNY publishes "Tons" without specifying short or metric. US short tons assumed (×0.90718). If metric, every NYC figure here is ~10% low.
The card above is not a curiosity. Of the 442 SDG indicators with usable data, the United States reports nothing at all for 130 — and on those, NYC vs the US is not a weaker comparison than NYC vs the world. It is the only one available, and the US is not in it.
After removing the indicators whose subject is a country rather than a place — “extent to which countries have laws and regulations that guarantee…” is not a quantity a city can hold a value for — 39 remain. Twelve of those are one coherent group, and every one has a DSNY or DEP series sitting opposite it.
| Indicator | Countries reporting | Years | NYC series that could fill it | Status |
|---|---|---|---|---|
| Municipal waste collectedEN_MWT_COLLV | 133 | 2000–2023 | Recycling Diversion and Capture Rates match 0.633 · proposed, not verified |
Graded & charted above |
| Hazardous waste generated per unit of GDPEN_HAZ_GENGDP | 107 | 2000–2024 | Wastewater Co-digestion and Biogas-to-grid Performance Indicators match 0.531 · proposed, not verified |
Ungraded proposal |
| Hazardous waste exportedEN_HAZ_EXP | 97 | 2000–2024 | DSNY Disposal Sites Used by Facilities by Year match 0.521 · proposed, not verified |
Ungraded proposal |
| Total waste generationEN_TWT_GENV | 97 | 2000–2023 | DSNY Waste Characterization - Comparative Results match 0.515 · proposed, not verified |
Ungraded proposal |
| Total wastewater treatedEN_WWT_TREAT | 94 | 2022level only | Watershed Water Quality - Wastewater match 0.542 · proposed, not verified |
Ungraded proposal |
| Hazardous waste treated or disposedEN_HAZ_TRTDISV | 93 | 2000–2024 | Dewatered Solids and Biosolids Allocations per Biosolids Management Practice Site match 0.51 · proposed, not verified |
Ungraded proposal |
| Proportion of hazardous waste that is treated or disposedEN_HAZ_TRTDISR | 88 | 2000–2024 | Dewatered Solids and Biosolids Allocations per Biosolids Management Practice Site match 0.506 · proposed, not verified |
Ungraded proposal |
| Hazardous waste importedEN_HAZ_IMP | 85 | 2000–2024 | DSNY Disposal Sites Used by Facilities by Year match 0.492 · proposed, not verified |
Ungraded proposal |
| Proportion of groundwater bodies with good ambient water qualityEN_H2O_GRAMBQ | 84 | 2017–2023 | Watershed Water Quality - Limnology match 0.516 · proposed, not verified |
Ungraded proposal |
| Total wastewater generatedEN_WWT_GEN | 84 | 2022level only | Wastewater Co-digestion and Biogas-to-grid Performance Indicators match 0.586 · proposed, not verified |
Ungraded proposal |
| Proportion of wastewater treatedEN_WWT_TREATR | 72 | 2022level only | Dewatered Solids and Biosolids Allocations per Biosolids Management Practice Site match 0.519 · proposed, not verified |
Ungraded proposal |
| Extent of human made wetlandsEN_WBE_HMWTL | 55 | 2021–2025 | NYC Wetlands match 0.497 · proposed, not verified |
Ungraded proposal |
Only municipal waste collected is graded and charted on this page. The other eleven are candidates an embedding matcher proposed; on hand-read lists its precision is roughly half. Under comparability spec v0.1 a grade is a human judgment, so none of them is a mapping until a person says so — which is the same rule that keeps seven pairs off the charts below.
3 of the twelve carry a single year. That supports a level comparison and not a trend — the same finding that removed road safety as this page’s headline demo. Not one of the twelve has a single United States datapoint; each was re-checked against the country observations rather than inherited from the screening proxy.
Source us-silent worksheet,
re-derivable with python3 probe/us_silent.py · reporting-country counts from a
sweep of every country and year for all 689 enumerated indicators
NYC candidates embedding search over all 2,400 NYC Open Data datasets · ungraded
Everything above runs city → UN: take an indicator, find the NYC series. That direction can only ever surface what the framework already asks about. Run it backwards — take every dataset 45 cities publish, find each one’s nearest indicator among all 519 named ones, and look at what is left over.
For 35 of those election datasets, the closest concept in the entire SDG framework is municipal waste collected — the indicator charted two sections above.
SG_SCP_PROCN, _HS and _LS sit in the goal tree
carrying no name and zero observations for zero countries.
The one adjacent indicator that does hold data asks whether a country has an action plan, and
its value is 1 for all 75 countries that answer.This is the broadest of the three categories — 28 cities, against 23 for elections and 8 for records access — and the least concentrated: New York is only 24% of it, where Milan is 66% of the election datasets. It is the one least likely to be an artefact of who happens to publish.
Two of the three are dominated by one publisher — Milan is 66% of the election datasets, Chicago 64% of the request logs. Procurement is the exception, spread across 28 cities with no publisher above a quarter, which is why it is the sturdiest of the three. One city publishing three hundred files is a filing habit. Twenty-eight cities each publishing some is a category of municipal government, and that is the number these claims rest on.
This counts titles that carry the vocabulary, in six languages, not datasets an embedding grouped together. An earlier clustering run put records access in 15 cities; only 8 of them publish a dataset that says so, and the rest were things like Media Releases that merely clustered nearby. The smaller number is the one that survives being checked.
Contract volumes NYC Government Databook (55,806 contracts, $147bn) · Portland OCDS
database, queried directly (9,913 contracts; $9.7bn is a floor — only 4,783 of them
carry a value field)
Source python3 probe/category_gaps.py ·
17,162 datasets from 45 city portals · framework searched across
all 519 named SDG base indicators · full method in the
inverse crosswalk
A blocked pair is not a failure of the pipeline. 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.
| Indicator | Verdict | Why it fails |
|---|---|---|
| Maternal mortalitySDG 3.1.1 | Blocked | The UN rate is per 100,000 live births, not per population — the denominator lives in a different NYC dataset. NYC also publishes pregnancy-associated deaths, a deliberately broader definition. |
| Child mortalitySDG 3.2.1 | Context only | UN counts deaths under five; NYC's series is under one. Both are raw counts, so the units "agree" — which is exactly the trap. The chart would look perfect and compare different populations. |
| PovertySDG 1.2.1 | Context only | The UN series is the international extreme-poverty line (~$2.15/day), near zero for any US city. NYC's own measure is an order of magnitude higher. Two numbers sharing a name and measuring nothing alike. |
| Deaths from air pollutionSDG 3.9.1 | Context only | UN reports a national death count from ambient air pollution; NYC reports an age-adjusted rate for ages 30+ from PM2.5 specifically. Different measure, age base and pollutant scope. |
| Inadequate housingSDG 11.1.1 | Context only | UN measures the share of urban population in slums. NYC's nearest analogue is a count of hazardous housing violations — an enforcement output that tracks inspection rates as much as conditions. |
| Built-up area per capitaSDG 11.7.1 | No NYC source | Tier 1 UN data with city-level slices across countries, 2000–2025. No NYC equivalent identified yet. Recorded so the gap stays visible rather than forgotten. |
| Safely managed drinking waterSDG 6.1.1 | No signal | Sits at ~100% for the US across the whole window, as it would for NYC. Two flat lines at 100 tell an analyst nothing. Excluded deliberately — a ceiling effect is a reason not to build a view, and that judgment belongs written down. |
We expected the UN to be the rich side and NYC the patchy one. On the city's flagship street-safety programme it is the other way round — and the consequence is not that the comparison fails, but that it only works in one direction.
So NYC can be placed and cannot be tracked. We can say the city sits 18th of 196, beside Spain and the Netherlands. We cannot say whether it is closing on them, because the world has one snapshot and NYC has eleven years.
That asymmetry is the argument for city data flowing toward the UN system rather than only the other way. The cities already hold the series. What is missing is the route by which it reaches the statistics everyone else compares against.