Public-Service Provider Directory ETL (HUD, HRSA, VA, OSM)

    1

    This kit is the fetch, transform, check and publish pipeline that keeps those straight, from a directory of 28,000 locations across all 50 states.

    $29

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Public-Service Provider Directory ETL (HUD, HRSA, VA, OSM)

    Example session with this skill installed

    We're building a find-help map for people facing homelessness in the US. Pull in health centers, housing authorities, housing counselors, VA facilities and mapped shelters for all 50 states, label where each record came from, and don't let a refresh silently drop a state.

    • Read your context and instructions
    • Compiled the public-service provider directory

    Summary

    Fetched five sources, assigned counties locally, excluded sensitive locations, and published 50 state files plus an index with source labels, licence and attribution. One source could not be fetched and was carried forward visibly instead of dropped.

    Fetch (node scripts/fetch-sources.mjs <source>)

    SourceResult
    hrsa17,788 active service-delivery sites (CSV, 42 MB)
    hud-pha3,449 housing authorities over 4 ArcGIS pages
    hud-hcs1,294 counseling agencies, 50+ states
    osm4,210 amenity=social_facility elements via Overpass
    censuscounty boundary zip, unzipped to work/census-counties-2025-20m/
    vano API key on this machine: skipped

    Build (node scripts/build-directory.mjs --work work --out public/data/providers --carry-forward va)

    • Counties assigned to 24,117 OSM and HUD records by point-in-polygon; HRSA rows kept their FIPS.
    • 291 OSM features excluded: domestic-violence or sexual-assault services by tag, name or operator; private access.
    • Coverage: all 50 states present; every source above its floor.
    • index.json: sources with kind and counts, OSM licence ODbL with attribution text and link, carriedForwardSources: ["va"] with the VA records' original dataUpdatedAt kept, and the disclaimer "not real-time vacancy ... call or verify before traveling".

    Output

    {
      "total": 28228,
      "bySource": { "hrsa": 17788, "hud-hcs": 1294, "hud-pha": 3449, "osm": 3919, "va": 1778 },
      "carriedForwardSources": ["va"],
      "states": 50
    }
    

    Tests

    node --test scripts/etl.test.mjs: 8 passed (synthetic shapefile, fake fetchers, coverage refusals, carry-forward).

    Connects securely to your tools. The creator never sees your data.

    About this skill

    A directory of places people can go for help is only as good as its honesty: which records are official and which are community-mapped, what licence they carry, whether a state quietly vanished in the last refresh, and whether a shelter for people fleeing abuse was republished with its coordinates. This kit is the fetch, transform, check and publish pipeline that keeps those straight, from a directory of 28,000 locations across all 50 states.

    Fetch (fetch-sources.mjs): HRSA health-center CSV, HUD public housing authorities (ArcGIS paging), HUD housing counselors, VA facilities (key from a file, camelCase pagination), OpenStreetMap social facilities via Overpass, Census county boundaries; each refuses a response below its minimum count or state coverage, the check that caught a silently truncated API.

    Transform (provider-etl.mjs): RFC 4180 CSV parsing, phone and website normalization, one record shape for every source with source and sourceUrl, categories, services and availability text that says verify before traveling; domestic-violence and private-access locations excluded; counties assigned by a grid-indexed point-in-polygon with holes.

    Shapefile reader (shapefile-lite.mjs): reads the Census .shp and .dbf directly, no package.

    Publish (build-directory.mjs): one JSON per state and an index naming every source, its kind (official-federal, community-mapped, provider-submitted), counts, the ODbL licence and attribution text, carried-forward sources when an optional fetch failed, and the disclaimer. Coverage of all 50 states and per-source floors asserted; a missing required source fails the build.

    reference/sources-and-licensing.md: the source table, why 211 is linked not imported, the exclusion rules, drift expectations, and the lessons behind each check. 8 offline tests with a synthetic shapefile and fake fetchers. Node 20+, no packages.

    U.S. only. No 211 data. Node 20+ required. Shapefile reader supports polygon types only. VA source requires an optional API key.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 19 days ago

    Requires

    Node.js 20+
    unzip or another zip tool

    Permissions

    Read Files
    Terminal / Shell
    Write Files
    Network Access

    Shell to run the node scripts and tests; read files for the fetched sources and the API key file; write files for work/ downloads and the published state files and index. Network access only in fetch-sources.mjs to data.hrsa.gov, egis.hud.gov, data.hud.gov, the VA API host, overpass-api.de and www2.census.gov. No environment variables.

    Listed19 days ago

    What's inside

    Frequently Asked Questions

    This skill provides a pre-built, zero-dependency Node.js pipeline that automates the collection, cleaning, and geographic tagging of official U.S. public service data, saving you from manual scraping or expensive data vendor fees.

    The package includes the full ETL source code, built-in validation tests, and a specialized Census shapefile reader, allowing you to generate state-specific JSON files for immediate use in your applications.

    Yes, the pipeline includes dedicated compliance tracking that automatically labels data sources and manages ODbL attribution requirements for OpenStreetMap data to ensure your final directory is legally compliant.

    This is a standalone Node.js 20+ implementation; it generates standardized JSON outputs that can be consumed by any AI agent, LLM-based application, or frontend framework that supports local data ingestion.

    While it currently supports health centers, housing authorities, and VA facilities, the modular architecture allows you to add custom sources as long as they provide geographic coordinates or valid addresses.

    Updates are released periodically to accommodate changes in official government API structures (like HRSA or HUD) and to ensure the Census boundary readers remain compatible with the latest administrative updates.

    New in Data & Databases