US metros medical and fire-rescue readiness

Brief: US metros medical and fire-rescue readiness

This is the prompt as it was run on 31 October 2026. The operator's own words were: "use all the data and MCP to pick top 10 metros of US by population and see if the council is ready, adequate for medical services and fire rescue services, measure and tell me if they are good and generate HTML report at the end, use data from digital atlas only and then rank them", followed by the meta-prompt below. Mid-run the operator added "use it internally here only, we have full access to data", which authorised running the published actions in-process on the production data when the public gateway was returning 502s (see steps.md).

Role

You are an urban services analyst with exclusive access to Digital Atlas via MCP.

Hard constraints (non-negotiable)

Use only Digital Atlas / Propheus MCP tools and data. Zero external web searches, knowledge, or other sources for any number, ranking, or count.
All population figures, place counts, categories, and metrics must come from Digital Atlas tool responses.
If a required metric or category does not exist in Digital Atlas, state that clearly and do not invent substitutes.
Prefer the latest served release for every metric. Always record the release and source vintage returned by the tools.

Objective

Identify the top 10 US Metropolitan Statistical Areas (CBSAs / metros) by total population using only Digital Atlas.
For each metro, measure the adequacy / readiness of medical services and fire rescue services using only place inventories, population, and related Digital Atlas metrics.
Create a transparent composite readiness score, rank the 10 metros, and judge whether each is "Good", "Adequate", "Strained", or "Inadequate".
Produce a single, self-contained, well-styled HTML report as the final deliverable.

Required workflow

Phase 1: discover capabilities (do this first)

Call the catalog tools to list layers, families, and place categories.
Explicitly search for place categories and metrics related to: hospital, medical center, emergency room, clinic, urgent care, fire station, fire department, emergency services, ambulance, EMS.
Note the exact category names (L1/L2) and any anchor types (e.g. "hospital") that Digital Atlas actually supports.
Confirm how CBSAs are addressed (entity type, GEOID format, context resolution).

Phase 2: rank metros by population

Resolve the largest CBSAs by population using Digital Atlas geography and population tools (area_profile.population or equivalent people-layer metrics).
Extract the top 10 by total population. Record official name, CBSA code/GEOID, and exact population figure plus release for each.

Phase 3: measure services for each of the 10 metros

For every metro:

Resolve a proper context (or equivalent geography handle) covering the full CBSA.
Pull total population (and, if available, daytime population or relevant age cohorts).
Count relevant places:
- Medical: hospitals, medical centers, emergency rooms, major clinics, etc. (use exact categories / anchors discovered in Phase 1). Prefer the "anchors" endpoint when it lists hospitals.
- Fire rescue: fire stations / fire departments / emergency services facilities (whatever categories exist).
Compute at minimum: absolute counts; facilities per 100,000 residents; any available density, coverage, or accessibility metrics Digital Atlas can return.
Note data quality / coverage notes returned by the tools.

Phase 4: scoring and ranking

Define a transparent composite score (show the formula in the report): a medical readiness component, a fire rescue readiness component, and optional supporting signals only if they come from Digital Atlas. Normalize within the group of 10 (or against national percentiles if the tools provide them). Assign qualitative labels: Good, Adequate, Strained, Inadequate. Rank the 10 metros from strongest to weakest overall readiness. Break ties clearly.

Phase 5: HTML report

Generate one complete, self-contained HTML file with: clean, professional, mobile-friendly styling; an executive summary with the final ranking; a methodology section (exact Digital Atlas tools, categories, releases, and scoring formula used); a ranked table of the top 10 with population, medical metrics, fire metrics, composite score, and qualitative rating; individual metro cards with key numbers and short interpretation; a data-quality and limitations note; a footer with generation timestamp and "Data exclusively from Digital Atlas via MCP". Do not embed external images or scripts. Pure HTML and CSS.

Output rules

Show your work: list the key tool calls and the exact numbers returned.
Never hallucinate a facility count or population figure.
If fire stations cannot be reliably counted from available categories, say so and score only what is measurable.
Final answer must include the full HTML report (or a clear instruction that the HTML is ready to copy).

Start instruction

Begin by exploring the Digital Atlas catalog for relevant place categories and population metrics.

What the brief got wrong, learned in the run

Fire stations are not a category. The only match is the level-2 category "Fire, Police & Emergency Services", which mixes fire stations, police and sheriff stations, and ambulance and rescue squads. The run pulled every listing of that category and classified it by name; the rules and the unclassified share are in report.md.
No action ranks metros by population. find_geography.search finds CBSAs by name and area_profile.population gives one context's total. The run ranked nationally by summing the served people layer's tract population by CBSA in the same process, then verified each of the ten with the published actions and with metrics.compare on a comparison set. A future brief should name the candidate CBSA codes as input.
Whole-CBSA populations need missing_values: exclude. New York has 14 of 4,928 tracts with no published value; under the default withhold rule the metro returns no total at all.
No emergency-room, medical-center or clinic categories exist, and no capacity, response-time or coverage metric exists at metro scale. Readiness here means facility listings per resident.

What this case shows can be done

Whole metros as one geography. context.resolve with census:cbsa:<code> gives a handle over every tract of a CBSA (up to 4,928 for New York) in under 50 ms, and area_profile.population and places.count work on it directly. The same works for counties, states, places and ZCTAs.
Exact per-capita counts by category. places.count with a context, filter.categories and per_capita: true returns open places per resident for any of the 314 level-2 categories, with closed places reported separately. Any facility-density question (grocery, pharmacies, schools, banks, gyms) is the same call with a different category.
A national anchor for free. The same count without a context gives the national rate from the same release, so a relative ranking can be shown beside an absolute benchmark.
A mixed category can be split by name. Where the taxonomy lumps things together, places.find paged by cursor returns every listing, and a documented name rule separates them; the unclassified share tells the reader how much to trust the split.
The tool can rank the candidates itself. metrics.compare on a comparison-set context orders the handles on a metric, which replaces outside knowledge once the candidates are named.

What it suggests next

The same brief for the next ten metros, for the counties inside one metro, or for one state's metros, with the CBSA codes given as input; about 50 public MCP calls for ten geographies.
Other readiness questions with the same shape: primary care, pharmacies, dialysis centres, nursing and assisted living per older resident (age cohorts are in the people layer), grocery per resident for food access, schools per child.
A change view: the places release is a snapshot, so a second run on a later release gives the direction of change; area_change cannot supply it for place categories today.
What it cannot do yet: response times, drive-time coverage of a whole metro (routed catchments are point-based and capped at 30 minutes), beds, staffing or apparatus. A brief that needs those should say so up front.