Skip to content

diffbot/diffbot-skills / skills/diffbot-places

diffbot-places

Unverified

Cities, counties, subregions, states, provinces, countries, and points of interest in the Diffbot Knowledge Graph by name, population, prominence, containing place, or proximity. MUST USE skill when the answer is a list of places or a fact about one (i.e. population), rather than recalling from memory. Triggers on: list cities, list countries, largest cities, biggest city in, population of, how many people live in, places in, cities near, states in, provinces, counties, regions, points of interest, landmarks.

Diffbot Place Search

Find geographic entities in the Diffbot Knowledge Graph: cities, counties/subregions, states/provinces/regions, countries, and points of interest. This is type:Place DQL with the place hierarchy, the right sort keys, and the known gaps already mapped.

Sibling note: /diffbot-people for people, /diffbot-organizations for companies located somewhere.

Sibling skills: /diffbot-organizations (companies, including those located somewhere), /diffbot-news, /diffbot-deals. Use /diffbot-dql for anything outside these shapes.

Workflow

Step 1 — check the CLI, then db dql init

~/.diffbot/venv/bin/db --version
~/.diffbot/venv/bin/db dql init

The db CLI is a one-time user install at ~/.diffbot/venv/bin/db, the same way gh or wrangler is — this skill never installs, upgrades, or uninstalls it. If --version fails with No such file or directory, stop and show the user this line to run or approve themselves (it is deliberately not pre-approved):

python3 -m venv ~/.diffbot/venv && ~/.diffbot/venv/bin/pip install 'diffbot>=3.1.1'

If it prints a version below 3.1.1, fails with No such option '--version' (an old diffbot-python install, which predates the flag), or warns about the diffbot-python rename, point the user at Upgrading the db CLI in the plugin README and stop. Never run an install, upgrade, or uninstall line yourself.

Save output files under .diffbot/tmp/ in the working directory. db creates that folder itself, so don't mkdir it first. db, jq, and cat are not pre-approved — Claude will ask before running them.

init refreshes ~/.diffbot/ontology.json and verifies a token (DIFFBOT_API_TOKEN env var, else ~/.diffbot/credentials). If neither exists the user must run:

echo "DIFFBOT_API_TOKEN=YOUR_TOKEN_HERE" > ~/.diffbot/credentials && chmod 600 ~/.diffbot/credentials

Tokens: https://app.diffbot.com/get-started/. Never echo the token.

Step 2 — pick the right type

Query the narrowest type that fits. type:Place matches everything including 12M points of interest; the subtypes are far more selective.

TypeApprox. countCovers
type:Country634Countries
type:Region1,671States, provinces, top-level subdivisions
type:Subregion28,160Counties, districts, second-level subdivisions
type:City3.7MCities, towns, municipalities
type:AdministrativeArea6.5MSuperset of the four above
type:Place18.5MEverything, including POIs

Points of interest — parks, landmarks, venues — are Place records that are not administrative areas. There is no Landmark type in practice (type:Landmark returns zero):

type:Place not(types:"AdministrativeArea") descriptors:"national park"

placeType is an equivalent filter to the type name (placeType:"Country" ≡ type:Country); POIs carry placeType:"other".

Step 3 — pick the levers

Containment — location or isPartOf

Two fields do this, and they are near-equivalent as filters:

type:City location.country.name:"Japan"
type:City isPartOf.name:"France"
type:Subregion location.region.name:"California"
type:Region location.country.name:"United States"

location fields: .city.name, .subregion.name, .region.name, .country.name, .metroArea.name, .latitude, .longitude, .postalCode.

isPartOf is the full containment chain — for Lyon it reads Metropolitan Lyon < Rhône < France < Metropolis of Lyon < Arrondissement of Lyon. Measured against location, they agree closely: French cities are 28,361 via isPartOf and 28,353 via location.country.name, and both return exactly 46 once population>100000 is added.

Prefer location.* when you want a specific administrative level (country vs region vs city), and isPartOf when you want "contained by X at any level" or need the chain itself. isPartOf is absent from the default JSON payload — request it with get: or --spec, or it will look empty. See the payload note in Step 5.

Ranking

FieldTypeUse
populationIntegerpopulation>1000000, revSortBy:population
importanceFloat (0–100)Prominence score. Already reflected in the default ranking — sort by it only to override a different sort
wikipediaPageviewsIntegerAttention proxy; also wikipediaPageviewsLastQuarter, …LastYear, and …Growth variants
areaInteger
nbIncomingEdgesIntegerHow connected the entity is in the KG

Leave the sort off unless the user asked for an ordering. The default ranking already bakes in relevance and prominence, and an explicit sort overrides it — usually for the worse. Measured: homepageUri:"openai.com" unsorted returns OpenAI, while revSortBy:nbEmployees returns "OpenAI for Developers"; anthropic.com unsorted returns Anthropic, sorted returns "Claude Builder Club".

Add a sort only when the ordering is the question — "largest cities by population", "most recent rounds", "biggest acquisitions". (Articles are the exception: see /diffbot-news, where sortBy:date is the default.)

For places specifically: strict:name:"Springfield" unsorted and with revSortBy:population return an identical top three (Missouri 169k, Massachusetts 156k, Illinois 114k) — the default already surfaces the prominent one.

Proximity

near(...) resolves a single anchor entity (the first match) and filters by distance; default radius 15km, override with mi or km:

type:City near(type:Place name:"Paris", 50km)
type:Place not(types:"AdministrativeArea") near(type:Place name:"Yosemite", 20mi)

Other fields

name, allNames, description, allDescriptions, descriptors, summary, postalCodes, areaCodes, headOfPlace (mayors, governors, heads of state), image.

Worked examples

type:City location.country.name:"Japan" population>1000000 revSortBy:population
type:City location.country.name:"Germany" revSortBy:population
type:Subregion location.region.name:"California" revSortBy:population
type:Region location.country.name:"United States"
type:City near(type:Place name:"Paris", 50km) revSortBy:population
type:Place not(types:"AdministrativeArea") descriptors:"national park"

Known gap — continents are not a field

There is no continent field. countryGroup exists but is populated on ~49 records, and an unrecognized path like countryGroup.name:"Europe" is silently ignored — it returns the unfiltered count, which looks like a working query. Watch for a hit count equal to the bare type: count; that is the tell.

For "all countries in Europe", match the prose description and verify:

type:Country description:"in Europe"

This returns ~89 hits for Europe: over-inclusive (countries merely mentioning Europe leak in) and it double-counts duplicate records. Export it, then filter the list yourself before presenting, and tell the user the list was text-matched rather than pulled from a continent field. With only 634 countries total, exporting all of type:Country and filtering against your own knowledge is also a legitimate approach.

Step 4 — probe before committing

~/.diffbot/venv/bin/db dql probe \
  'type:City location.country.name:"France" population>100000' \
  'type:City location.country.name:"France"' \
  'type:Place name:"Europe"'

Two failure signatures to watch for:

  • Hit count == the bare type: count → your filter path is wrong and was ignored.
  • Zero hits → the value is wrong for this field, or the concept isn't modelled at all (continents: isPartOf.name:"Europe" on countries returns 0).

probe fails the whole batch if any single variant is rejected by the API — fix the offending clause and re-run rather than assuming the others were checked.

Place names are matched with contains, so name:"Springfield" pulls in every Springfield on earth. Add strict: plus a location.country.name or location.region.name filter to pin one, and facet:placeType to see what mix of cities, subregions, and POIs a name spans.

Diacritics are not normalized, and coverage is inconsistent. Query both spellings in one shot — name:or("Kopavogur","Kópavogur") — whenever a name could carry an accent. Verified: name:"Kópavogur" returns 1 city and name:"Kopavogur" returns 0 (as do allNames: and a prefix), while type:Place name:"Reykjavik" and "Reykjavík" both return 44. A zero from the ASCII spelling alone is not evidence of absence — never report a place as missing from the KG until the accented form has also returned 0.

Step 5 — export and display

~/.diffbot/venv/bin/db dql export "<DQL>" --out .diffbot/tmp/places.csv \
  --spec "name,Place;population,Population;location.region.name,Region;location.country.name,Country;importance,Importance" \
  --size 50

The default JSON payload is not the full entity. A plain --format json export of a Place returns only name, description, placeType, importance, types, and provenance — population, location, and isPartOf are all absent, even though you can filter on them. Ask for them explicitly:

~/.diffbot/venv/bin/db dql export '<DQL> get:name,population,location,isPartOf' --out .diffbot/tmp/places.json --format json --size 50

--spec already does this for you, which is why the CSV above shows populations. Only raw JSON exports need get:. A null field in a JSON export usually means you didn't request it, not that the data is missing.

Display

  1. Render a markdown table with columns matched to the question — name, population, containing region/country.
  2. Deduplicate. The KG holds multiple records for the same place (e.g. two Tokyo rows with different populations — city proper vs. metro). Collapse same-name rows in the same parent and say which figure you kept, or show both labelled.
  3. Format populations with thousands separators; importance is a 0–100 prominence score, not a rank — only show it if it's relevant.
  4. Print the final DQL in a plain code block.
  5. Tell the user the saved file path, and offer more rows (--size N, --from K) or a refinement.
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 467kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 319kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 319kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 319kUpdated Claude CodeCodexVerifiedInstall