A data dictionary for streaming availability research

Abstract. This reference defines every field surfaced by PlatformTrack — on country pages, platform pages, the trending bands and the underlying database — and specifies how each field is computed, what it includes and what it deliberately excludes. It is intended as a companion citation document for researchers who want to be precise about which figure they are reporting.

1. Why a data dictionary matters

Streaming-availability journalism and research is full of ambiguous numbers. “Netflix has 8,000 titles in Turkey” can mean a half-dozen different things depending on whether you are counting unique titles or availability records, whether you are including transactional rentals, whether you are date-anchored, and whether your sample is the full catalog or a popularity window. This document removes that ambiguity for everything PlatformTrack publishes.

Every figure on PlatformTrack maps to a field defined below. Every field has a single, stable definition that does not change between sync runs without a documented version bump.

2. Core entities

Platform

A streaming service tracked by the index. Each platform has a TMDB watch-provider ID, a human-readable name, a slug, a logo, and a platform type:

• subscription — SVOD: included with a monthly fee (Netflix, Disney+, Max).
• buy — transactional purchase or rental (Apple iTunes, Google Play).
• ads — ad-supported free streaming (Tubi).
• free — free streaming without registration (rare).
• rent — rental-only catalogs (less common standalone).

The current platform list spans 17 services across these categories. Researchers should be deliberate about which platform types they include in a given analysis — aggregating subscription and transactional catalogs into a single “total available” number is rarely the right move.

Country

A national market the index tracks. Each country has an ISO 3166-1 alpha-2 code, a slug, a name, and a region grouping (Europe, Middle East, Asia, Africa, Americas, Oceania). Regions are used for cross-sectional grouping on the index leaderboards and have no operational meaning beyond presentation.

Availability record

The core table row of the index. One availability record exists per (platform, country) pair where the platform is present in that country’s TMDB watch-provider data. Each record carries the catalog-size fields defined below and a last_synced timestamp.

3. Catalog-size fields

Four count fields are recorded per availability record at every weekly sync.

title_count

The total number of titles — movies plus series — available on the platform in the country. This is the headline number on every platform-country page.

Computed as movie_count + series_count, where both components are sourced from the total_results field of TMDB’s Discover endpoint. Includes every title TMDB lists as available under the platform’s watch-provider ID in that region, irrespective of subscription tier, stream quality or localization status.

Does not include: titles available only via affiliate channels of the platform that TMDB lists as separate providers (e.g., a Netflix Basic with Ads tier that surfaces as a distinct provider ID in some markets).

movie_count

The number of movies (TMDB media_type=movie) available on the platform in the country. Pulled from total_results on the Discover movie endpoint.

series_count

The number of television series (TMDB media_type=tv) available on the platform in the country. Pulled from total_results on the Discover tv endpoint. Counts series as one record regardless of season or episode count.

exclusive_count

The number of titles in the popularity sample for this (platform, country) pair that do not appear in any other platform’s popularity sample for the same country. See §4.

This is a sample-conditioned measure. A platform with structurally exclusive but unpopular catalog items will under-count relative to the structural truth.

4. Popularity sample

Every sync harvests the top 40 (typically) most-popular titles per (platform, country) pair, sorted by TMDB’s popularity field. These power the trending bands on the homepage, the “top titles” lists on platform-country pages, and the exclusivity calculation.

Each sample row carries:

• tmdb_id — the canonical TMDB title identifier.
• title — the title as TMDB records it (English by default; locale-aware variants are not stored).
• media_type — movie or tv.
• popularity — TMDB’s popularity score at sync time. Researchers should treat this as a relative signal within a sync, not an absolute cross-sync metric (the underlying formula is opaque and changes).
• vote_average — the TMDB user-vote average. Stable across syncs.
• poster_path — relative path for the title’s poster artwork.
• is_exclusive — boolean flag, recomputed per sync per §3.
• rank_pos — the title’s rank within the platform-country’s popularity sample.

5. Derived measures

Reach

For a platform, reach is the count of distinct countries in which it has at least one availability record with title_count > 0. Reach maxes out at the country count of the index (57 at time of writing).

Streaming depth

For a country, streaming depth is the sum of title_count across all platforms with at least one availability record in that country. It is the headline metric of the country leaderboard. A country with three deep catalogs will outrank one with seven shallow catalogs.

Movie/series ratio

Computed as movie_count / series_count at the availability-record level, or aggregated to the country or platform level. Useful for identifying platforms or markets with structurally different content mixes — e.g., Crunchyroll’s series-dominant catalog versus Apple iTunes’ movie-dominant rental catalog.

6. Snapshot table

A separate pt_snapshots table preserves title_count, movie_count and series_count at the (platform, country, date) level. One row per pair per weekly sync. This is the canonical store for historical analysis and time-series queries.

The live availability table is overwritten weekly; the snapshot table is append-only and is the source of truth for any “has Netflix Spain grown or shrunk?” analysis.

7. Computed fields surfaced on pages

Several figures shown on PlatformTrack pages are computed at render time from the core fields. They are documented here for transparency.

• Average catalog per country (on platform pages) — total_titles / country_count across all availability records for that platform with title_count > 0.
• Top platform per country (on the ranking page) — the platform with the maximum title_count in that country.
• Largest market (on platform-ranking page) — the country with the maximum title_count for that platform.
• Catalog composition (donut on platform-country pages) — movie_count / title_count and series_count / title_count.
• Catalog trend (sparkline on platform-country pages) — up to twelve most-recent snapshots of title_count from pt_snapshots.

8. Fields not in the index

Several measures researchers commonly ask about are not in PlatformTrack:

• Hours watched, subscriber counts, ARPU. Out of scope. PlatformTrack measures supply, not demand.
• Stream quality (4K vs HD). Not differentiated.
• Localization (dubs, subtitles). Not tracked.
• Title-level entry/exit timestamps. The popularity sample is the only title-resolution we maintain, and it is overwritten weekly. Full-catalog entry/exit tracking would require sampling the entire catalog at each sync, which is operationally out of scope.
• Pricing. Subscription pricing varies by market, tier and promotion; PlatformTrack does not record it.

9. Versioning

This data dictionary versions independently of the methodology document. Material changes — new fields added, definitions clarified, or measures retired — are logged inline in this document. The current version date is the publication date of this page (see footer).

For citation, use the “How to cite” block below. For corrections or precision requests, write to [email protected].