Cell Tower Gis Data By State Data: Dataset Structure and Field Coverage

There is no single federal file called “US towers by state.” A workable GIS build stacks three layers that answer different questions:

1. Regulatory points (national). The FCC's Antenna Structure Registration files carry coordinates, heights and owner entities for every registered structure - about 8.7 million rows across nine .dat files inside a 208 MB a_tower.zip, covering 1.09 million registration determinations back to 1990.

2. Observed radio cells (global, partitioned by country/MCC). OpenCellID refreshes daily CSV.gz dumps with radio, mcc, cell, lat/lon, range and sample counts for millions of logical cells over a rolling 18-month window.

3. State and municipal GIS layers (patchwork). Data.gov's discovery layer indexes 46 results for the query “cell tower,” including Austin and Baton Rouge tower points, Los Angeles small cells, San Francisco wireless facilities and Connecticut Siting Council sites - published in json, csv, kml, kmz and geojson.

Only the third layer is genuinely “by state” as published; the first two are national or global point sets you filter by state afterward. That distinction decides where each belongs in your workflow.

Coverage is a patchwork, not a wall-to-wall map. In the Data.gov Cell Tower Datasets Catalog, the observed state and city coverage concentrates in four places: Texas (Austin) and Louisiana (Baton Rouge) publish cell tower point layers; California contributes both Los Angeles small cells and San Francisco wireless facilities; and Connecticut's Siting Council publishes its approved tower sites as public-domain records.

A second query widens the net. The Data.gov Antenna Tower Datasets Catalog returns 8 federal, state, county and city inventories, adding Loudoun County, Virginia telecom tower sites alongside San Francisco, Los Angeles and the Connecticut Siting Council - so observed coverage spans California, Connecticut and Virginia on the antenna side.

Two practical consequences follow. First, states that do not publish their own layers are not empty - they simply fall back to the FCC registry and OpenCellID observations. Second, because Data.gov re-harvests agency metadata daily while individual constituent datasets were last modified anywhere between 2019 and August 2026, check each layer's own date before treating it as current.

The FCC route gives every state a consistent layer even where no local GIS office has published one. The FCC Data Hub ships a complete Antenna Structure Registration snapshot weekly plus change files each weekday: a_tower.zip is 208 MB compressed with roughly 8.7M rows across nine .dat files, d_tower.zip adds 3.4M determination rows at 56 MB, and l_cell.zip weighs 24.8 MB. Registrations reach back to 1990, and no key or registration is required.

The walkthrough most teams converge on:

Because the extract is U.S. Government commercial delivery terms under 17 USC 105, the resulting state layers can be redistributed commercially without attribution - unusual among the sources here.

TowerMaps is the only consolidated option in this slice: more than 600,000 antenna-site locations across all states, compiled from several hundred site owners since 1997, delivered quarterly as dated issues in ESRI geodatabase or Excel format. Owner-level attributes per structure are its differentiator - neither the FCC extracts nor OpenCellID fields ship a comparable consolidated ownership table.

Sanity-check any per-state layer against national totals before styling it. The Wireless Infrastructure Association's annual “Wireless Infrastructure By the Numbers” report counts 158,500 purpose-built US towers and 254,850 macrocell sites at end-2025. Wireless Estimator's independently compiled list reaches roughly 149,196 towers across 276 tower-owning companies, with counts timestamped per company against a verification stamp of 7/1/2026.

Those two figures bracket what a purpose-built tower layer should contain nationally - a state map claiming far more is almost certainly counting antennas, small cells or registration actions rather than structures. The FCC file will legitimately exceed them because its 1.09M determinations include unbuild and terminated records alongside standing structures, which is why filtering on determination status matters before you publish a count.

Sizing context helps when a stakeholder asks why one state looks sparse: consolidation means the national footprint concentrates in a few portfolios, with Wireless Estimator ranking all 276 companies and republishing only its Top 10 with attribution.

The fastest free path combines three sources in one afternoon of work:

The output is a defensible multi-source state layer rather than a single-vendor black box, and every input remains free to re-pull on your own schedule.

The telecom tower REITs data guide is the pillar overview of all 22 pooled datasets in this industry, including the equity-market feeds for AMT, CCI and SBAC that sit outside a pure GIS workflow.