Radio Airplay Data for Music Analytics: What Is in It and What to Build

Streaming numbers tell you what listeners chose; airplay data tells you what gatekeepers chose — hundreds of programmers and DJs making editorial decisions daily, timestamped and geolocated. For analytics teams, that’s a distinct, underused signal. Here’s what radio airplay data contains, what it’s good for, and how to work with it programmatically.

What’s actually in an airplay dataset

The atomic unit is the spin: one recording, one station, one timestamp. Around it, a usable dataset layers recording identity (title, artist, ISRC, album, label, genre), station metadata (name, network/source, format, city/region/country, coordinates), and derived aggregates (per-track and per-station totals, trends, chart positions). The SpinCounts dataset spans 6,700+ U.S.-focused stations — FM streams including iHeartRadio, SiriusXM channels, and internet radio — collected continuously from station playout feeds; collection method and freshness are documented openly on the methodology page, which matters when you’re deciding whether to trust a feed enough to build on it.

Why analytics teams care about radio

It’s an editorial-signal series, not a consumption series. Every spin is a professional’s choice made under audience risk — which makes airplay a leading indicator of the push behind a record rather than the response to it. Cross-referencing the two answers questions neither answers alone: is this record radio-led or streaming-led? Is the label spending? Is regional heat organic?

It’s geographic by nature. Stations have coordinates; streams mostly don’t (at station-level granularity). Airplay maps expose market-by-market adoption — tour-routing intelligence for the industry, regional-trend detection for researchers.

It’s fast. Adds and rotation changes happen days or weeks before their downstream effects. A record jumping from 5 stations to 25 in a fortnight is a detectable event with a timestamp — the kind of discrete, datable signal that survives being put in a model.

Concrete builds: label roster dashboards joining spins to streaming and social series (the workflow half of this is covered in the label guide); A&R screens flagging multi-station adoption of unsigned artists; market researchers quantifying format trends; and quantitative teams using airplay as an alternative-data series on music-industry activity. One boundary we state plainly, because serious teams ask: the dataset contains no personal information and no material non-public information — it’s observed broadcast behavior, publicly verifiable on the site itself.

How to access it

Start free, no contact required: the SpinCounts Data API includes free authenticated access to the charts endpoints — live rankings by genre with movement — which is enough to prototype trend detection or a dashboard widget today. The full firehose — timestamped spins with complete recording and station profiles, filterable by artist, track, station, geography and time — runs on per-developer keys with straightforward REST semantics and OpenAPI docs. Pricing is contact-first because usage patterns vary wildly between a dashboard and a research pipeline; the docs, schemas and trial keys are self-serve at /developers/. And because every data point also exists as a public page, you can audit any API response against the site with a click — an unusual property in this category, and deliberately so.

Evaluating any airplay dataset (including ours)

Four due-diligence questions worth asking any vendor: What’s the coverage shape — which markets and formats, not just the station count? What’s the latency — minutes, hours, or weekly batches? Is the matching method documented — how do raw feed strings become recording-level rows, and what’s the junk-filtering story? And is the data independently verifiable, or does the API’s word have to be taken on faith? We publish our answers; insist on the same from anyone else.

Frequently asked questions