Capability

Geospatial

Data LiveAPI Planned

Every jurisdiction carries its real boundary, so one coordinate resolves to the complete stack of governments that touch it, down to the elementary school district, in a single query with no third-party district service anywhere in the path.

One point, every jurisdiction

This is an actual result from the production graph for a coordinate in downtown Phoenix, returning nine jurisdictions ordered smallest first in one round trip.

Phoenix Elementary District
school-district
15
Arizona State Senate District 11
state-legislative-senate
106
Arizona State House District 11
state-legislative-house
106
Phoenix Union High School District
school-district
174
Arizona's 3rd Congressional District
congressional-district
206
Phoenix city
municipality
519
Maricopa County
county
9,226
Arizona
state
113,985
United States
country
3,635,169

Area in square miles, logarithmicsq mi

Two details in that result separate a real spatial model from a simple district lookup. Both school districts appear, because Phoenix Elementary and Phoenix Union High School overlap the same address wherever elementary and secondary districts are drawn independently. The stack is also not a strict hierarchy, since the legislative districts are larger than the city without being contained by it. We return the true containment set ordered by area rather than pretending governance nests cleanly.

GET /v1/districts_by_point

{
  "point": { "lat": 33.4484, "lng": -112.0740 },
  "regions": [
    { "slug": "region:az-phoenix-elementary-district-0406300",
      "name": "Phoenix Elementary District", "level": "school-district",
      "area_sq_mi": 15, "vintage": 2024 },
    { "slug": "region:az-state-senate-district-11",
      "name": "Arizona State Senate District 11",
      "level": "state-legislative-senate", "area_sq_mi": 106 },
    { "slug": "region:az-phoenix-0455000",
      "name": "Phoenix city", "level": "municipality", "area_sq_mi": 519,
      "ocd_id": "ocd-division/country:us/state:az/place:phoenix" }
    // six more, ordered by area
  ]
}

Geometry is stored in PostGIS as SRID 4326 polygons behind a spatial index, and resolution is a single containment test ordered by area and filtered to the boundaries valid on the requested date. Coordinate input is exact, whereas address input is geocoded first, so a coordinate is always the more precise call.

Boundary coverage

The tiers below county are both the hardest to obtain and the hardest to keep current, which is where most of this work went.

Jurisdiction tier	Boundaries
Municipality	37,636
School district	13,317
State house	4,784
County	3,235
Precinct	2,256
State senate	1,964
Congressional district	469
Ward	285
State	56
Judicial district	40

Measured from the production graph on 2026-08-20.

What a region carries besides its shape

A boundary alone answers where something is, while these fields answer who lives there and how to reach them without bringing in a second data vendor.

Field	What it is
`population`	ACS population estimate for the jurisdiction
`vintage`	Census vintage year the boundary was drawn from
`timezone`	The zone that decides what "polls close at 7pm" actually means
`ocd_id`	Open Civic Data identifier, for joining to outside datasets
`languages`	Most-spoken languages at home, priority ordered, from ACS
`land_area_sqm` and `water_area_sqm`	Land and water area, for real density arithmetic
`fips` and `state_fips`	Census identifiers, for joining to any federal dataset

The languages field drives outreach. Colusa County, California returns ["es","en","de","vi"], so anyone planning ballot translation, canvass materials or a multilingual voter guide can see where to spend, attached to the same boundary they just resolved.

Boundaries are versioned

Districts get redrawn, so a boundary is never a fact about now but a fact about a window of time, which is why every region carries its Census vintage and the same validity dates as everything else in the graph.

Districts as of a past date

GET /v1/districts_by_point?lat=33.4484&lng=-112.0740&as_of=2022-06-01

Ask what district an address sat in during the last cycle and you get the answer that was true then rather than today's boundary applied retroactively, which matters a great deal when reconciling historic turnout or auditing a past targeting decision.

Derived boundaries

Some jurisdictions have no published boundary file at all, and state judicial districts are the clearest case as well as a documented gap for both major alternatives.

We build them ourselves. A judicial district is defined by statute as a set of whole counties, so we merge those counties into a single exact boundary and attach the court authority to it, which produces a real queryable shape for a jurisdiction that is otherwise invisible.

Derived boundaries are labelled as derived. A merged boundary carries the same provenance record as anything else and cites the statute it was built from, so you can always tell which shapes came from the Census and which we constructed.

Geography without geometry

Spatial queries are expensive, and most containment questions do not need one because the nesting is already stored in the graph as containment edges that you can traverse instead.

School districts in a county, and who is running

MATCH (county:Region { slug: 'region:az-maricopa-county' })
MATCH (county)-[:CONTAINS*1..2]->(sd:Region)
WHERE sd.level = 'school-district'
MATCH (o:Office)-[:SCOPED_TO]->(sd)
MATCH (c:Candidacy)-[:RUNS_IN]->(:Contest)-[:FILLS]->(o)
MATCH (c)-[:IS_CANDIDACY_OF]->(p:Person)
RETURN sd.name AS district, sd.population AS population,
       o.name AS office, collect(p.name) AS candidates

That single statement goes from a county to every school board race inside it. Resolving the same thing through a district lookup service and a separate candidate feed would mean a spatial call, a join on names, and a reconciliation problem you would own permanently.

Why this beats a district lookup service

Our assessment, not a vendor-confirmed comparison. The Civoren column is measured from the graph, while the left column describes the general shape of the address-to-district services we have evaluated.

Question	A typical district service	Civoren
Input	Address or ZIP	Coordinate, address, or a node slug
Levels returned	Federal and state legislative	Ten levels, down to precinct
School districts	Rarely	Elementary and secondary held apart
Precincts	No	Real boundaries
Judicial districts	No	Derived from statute
Historic boundaries	Current only	Versioned, query as of any date
Connected to races	A separate lookup you join yourself	Same graph, one traversal
Demographics attached	No	Population, languages, timezone, area
Raw geometry	Not exposed	GeoJSON on request