Every other analysis here runs city → UN: take an SDG indicator, find the municipal dataset that matches it. That can only discover what the framework already asks about. This runs it backwards — take every dataset a city publishes, find its nearest SDG indicator, and look at what is left over.

5,956 datasets from 11 city portals (language non-en), matched against all 519 enumerated SDG indicators — not the 442 with usable data, because a framework gap is a question about vocabulary rather than coverage.

What a result here means. A dataset far from every indicator means no SDG indicator's text is near this dataset's text. That is evidence about vocabulary, not proof of a conceptual gap — this project has already learned once that a null from the matcher is not evidence of absence. So the unit of evidence below is how many independent cities a theme appears in. A theme in thirty city catalogs is a category of municipal governance; a theme in one is that city's filing habit.

Positive controls

Datasets from the hand-verified NYC crosswalk. These are known to correspond to an SDG indicator, so they must land in the high-affinity region. If they fall in the tail, the tail is measuring retrieval failure and nothing below is trustworthy.

11 of 11 located · 0 fell in the tail.

Expected correspondence City Dataset Affinity Percentile In tail?
Population structure (demographics) Rostock Bevölkerungsstruktur 2005 0.633 91 no
Road traffic accidents (3.6.1) Madrid Accidentes de tráfico con implicación de b 0.535 90 no
Road traffic accidents and injuries (3.6.1) Milan Mobilità: incidenti stradali e persone inf 0.516 90 no
Municipal waste collected (11.6.1) Milan Rifiuti raccolti a Milano e nei comuni lim 0.489 88 no
Municipal waste collection (11.6.1) Belo Horizonte Coleta de Resíduos 0.474 79 no
Monthly road traffic accidents (3.6.1) Matera Bilancio mensile incidenti stradali anno 2 0.477 72 no
Air quality, hourly (11.6.2) Madrid Calidad del aire. Datos horarios desde 200 0.462 71 no
Water quality (6.3.2) Buenos Aires Calidad del Agua 0.418 61 no
Drinking water quality (6.1.1) Karlsruhe Jahresmittelwerte zur Trinkwasserqualität 0.408 57 no
Deaths from road traffic accidents (3.6.1) Fortaleza Número de Óbitos por Acidentes de Trânsito 0.333 44 no
Road traffic accidents with victims (3.6.1) Recife Acidentes de Trânsito com Vítimas 2016 0.394 32 no

What cities publish that the SDGs have no words for

The bottom 25% of each catalog by affinity — 1,516 datasets. Below are the title phrases that recur across that tail, counted by how many independent cities use them. These are not inferred categories; they are what the cities themselves called the data.

Cities Datasets Phrase

No phrase recurs across four cities. Expected for a run pooling five languages — Spanish, Italian, Portuguese, German and Croatian titles share essentially no bigrams, so this view is uninformative here by construction rather than because nothing recurs. The clusters below carry the result.

In detail

The same tail, clustered by embedding

A second view, kept because it groups datasets that share no vocabulary. k-means returns k clusters whether or not k themes exist, so each carries its coherence — the mean cosine of its members to its own centroid. Below 0.62, a cluster is a partition rather than a theme and is marked diffuse; 12 of 20 are. Read those as noise, not as findings.

Coherence Cities Datasets Terms Closest SDG indicator
0.895 6 241 elezioni, sezione, risultati, referendum, abrogativo Countries that have national urban policies
0.861 5 113 omi, quotazioni, immobiliari, semestre, locazione Performance index of data Infrastructure (Pi
0.711 3 71 orari, categoria, accessi, area, storica Countries that have national urban policies
0.698 5 33 telelavoro, personale, indeterminato, tempo, anno Proportion of project objectives of new deve
0.679 4 51 pirf, zeis Degree of implementation of integrated water
0.676 2 51 milanesi, accademico, universit, nelle, anno Open Data Inventory (ODIN) Coverage Index
0.644 7 39 covid- Number of new HIV infections per 1,000 uninf
0.633 6 69 ipem Universal health coverage (UHC) service cove
0.62 (diffuse) 2 26 geologia, sica, lote Total greenhouse gas emissions (including la
0.611 (diffuse) 10 69 (no term covers a fifth of this cluster) Average proportion of Terrestrial Key Biodiv
0.599 (diffuse) 7 11 kvaliteti, zraka, zagrebu, podaci, gradu Carbon dioxide emissions per unit of GDP at
0.594 (diffuse) 10 90 (no term covers a fifth of this cluster) Death rate due to road traffic injuries
0.585 (diffuse) 11 120 (no term covers a fifth of this cluster) Total public expenditure per capita on cultu
0.557 (diffuse) 9 72 (no term covers a fifth of this cluster) Proportion of municipal waste recycled
0.552 (diffuse) 7 70 licita, fortaleza Countries that have national urban policies
0.541 (diffuse) 7 69 zagreba, grada Proportion of countries with alignment of Na
0.535 (diffuse) 4 35 (no term covers a fifth of this cluster) Countries that have national urban policies
0.51 (diffuse) 7 34 (no term covers a fifth of this cluster) Countries that have legislative, administrat
0.503 (diffuse) 9 98 (no term covers a fifth of this cluster) Countries with procedures in law or policy f
0.469 (diffuse) 9 154 (no term covers a fifth of this cluster) Open Data Inventory (ODIN) Coverage Index

The coherent clusters in detail

elezioni, sezione, risultati, referendum, abrogativo — 241 datasets across 6 cities (coherence 0.895, mean affinity 0.294)

Most central to the cluster:

One per city, to show the spread:

omi, quotazioni, immobiliari, semestre, locazione — 113 datasets across 5 cities (coherence 0.861, mean affinity 0.32)

Most central to the cluster:

One per city, to show the spread:

orari, categoria, accessi, area, storica — 71 datasets across 3 cities (coherence 0.711, mean affinity 0.306)

Most central to the cluster:

One per city, to show the spread:

telelavoro, personale, indeterminato, tempo, anno — 33 datasets across 5 cities (coherence 0.698, mean affinity 0.316)

Most central to the cluster:

One per city, to show the spread:

pirf, zeis — 51 datasets across 4 cities (coherence 0.679, mean affinity 0.261)

Most central to the cluster:

One per city, to show the spread:

milanesi, accademico, universit, nelle, anno — 51 datasets across 2 cities (coherence 0.676, mean affinity 0.312)

Most central to the cluster:

One per city, to show the spread:

covid- — 39 datasets across 7 cities (coherence 0.644, mean affinity 0.307)

Most central to the cluster:

One per city, to show the spread:

ipem — 69 datasets across 6 cities (coherence 0.633, mean affinity 0.273)

Most central to the cluster:

One per city, to show the spread:

Reproducing

python3 probe/fetch_municipal.py
python3 probe/inverse.py