Data Sources

Where the data comes from

MyRigCheck aggregates game metadata from IGDB, system requirements from the Steam Store, hardware reference data from a curated 265-SKU set, and rig data from anonymized user-submitted rigs. The current dataset version is 1.0.0, released 2026-07-25. The full public export is at /api/dataset.json (CC BY 4.0).

Author & Editor: Ankur Halder

Last reviewed 2026-07-25 · Dataset version 1.0.0

Ingestion pipeline

The data pipeline runs in three phases. A weekly full re-ingest of every source rebuilds the read layer from scratch (correcting any drift in the previous week). Daily incremental updates patch in new games, review counts, and rig submissions. The "Data refreshed" timestamp in the site footer reflects the last successful build.

Extract — IGDB via the IGDB API; Steam Store via the public Storefront and SteamSpy-style estimates (no key required); the hardware reference is a static JSON set in the repo.
Transform — game records are joined against the hardware reference and the rig data to produce per-game compatibility statistics. System requirements from IGDB and Steam are merged with a "primary source wins" rule.
Load — the read layer is Mongo. The build tier (used during re-ingest) is rate-limited and write-protected; the runtime tier (used by the website) is read-only and cached.

Data sources

IGDB (Internet Game Database)

Game metadata

Game metadata, genres, themes, platforms, game modes, player perspectives, franchises, engines, and collections. The IGDB API is the primary source for taxonomy and the `igdb_*` fields on every game record.

Cadence: Weekly full re-ingest; daily incremental updates

Steam Store

System requirements, reviews, pricing, CCU

System requirements, peak concurrent users, review counts (positive + negative), pricing, and store metadata. The Steam store page is cited as a `sameAs` link in the schema for every Steam-listed game.

Cadence: Weekly

Hardware reference (data/benchmarks/)

Calibrated hardware scores

265 clean SKUs (CPUs and GPUs) with calibrated 0-100 hardware scores. JOINED into the read layer at build time, not seeded from a third-party dataset. The full methodology is in /methodology.

Cadence: Updated with each methodology version bump

User-submitted rig data

Rig coverage statistics

User rigs (CPU, GPU, RAM) submitted via the /check form. Stored locally + in an anonymous cookie. Used only for rig-coverage stats (e.g. "12% of submitted rigs can run this game at 60 fps"). Never sold, never shared with third parties, never published individually.

Cadence: Real-time, anonymized

Community FPS reports

Calibration data

Observed frame rates reported by users via the "Report FPS" button on every game page. Feeds the calibration loop for the FPS-estimate model. Per-rig data is never published.

Cadence: Continuous

Public dataset export

The full catalog (350,000+ games) plus the hardware reference is available as a single JSON file at /api/dataset.json. The file carries a top-level version field, a generatedAt timestamp, and a license field. License: CC BY 4.0 — attribution to MyRigCheck is required; no other restrictions.

The dataset is versioned semantically. Breaking changes to the schema bump MAJOR; additive fields bump MINOR; corrections and accuracy improvements bump PATCH. The changelog is at /CHANGES.md.

Inferences and derived data

Three derived fields are computed in the read layer and exposed on every game record:

popularity_score — composite of IGDB popularity, Steam review volume, and recent activity. Used to rank games on the home dashboard.
trend_score — week-over-week rank delta. Used to surface breakout games.
composite_rank — final ordering used by the sitemap, the home page, and the top-games prerender set. Combines popularity_score + trend_score with a freshness decay.

Where the data comes from

Ingestion pipeline

Extract — IGDB via the IGDB API; Steam Store via the public Storefront and SteamSpy-style estimates (no key required); the hardware reference is a static JSON set in the repo.

Transform — game records are joined against the hardware reference and the rig data to produce per-game compatibility statistics. System requirements from IGDB and Steam are merged with a "primary source wins" rule.

Load — the read layer is Mongo. The build tier (used during re-ingest) is rate-limited and write-protected; the runtime tier (used by the website) is read-only and cached.

Data sources

IGDB (Internet Game Database)

Game metadata

Cadence: Weekly full re-ingest; daily incremental updates

Steam Store

System requirements, reviews, pricing, CCU

Cadence: Weekly

Hardware reference (data/benchmarks/)

Calibrated hardware scores

265 clean SKUs (CPUs and GPUs) with calibrated 0-100 hardware scores. JOINED into the read layer at build time, not seeded from a third-party dataset. The full methodology is in /methodology.

Cadence: Updated with each methodology version bump

User-submitted rig data

Rig coverage statistics

Cadence: Real-time, anonymized

Community FPS reports

Calibration data

Observed frame rates reported by users via the "Report FPS" button on every game page. Feeds the calibration loop for the FPS-estimate model. Per-rig data is never published.

Cadence: Continuous

Public dataset export

The dataset is versioned semantically. Breaking changes to the schema bump MAJOR; additive fields bump MINOR; corrections and accuracy improvements bump PATCH. The changelog is at /CHANGES.md.

Inferences and derived data

Three derived fields are computed in the read layer and exposed on every game record:

popularity_score — composite of IGDB popularity, Steam review volume, and recent activity. Used to rank games on the home dashboard.

trend_score — week-over-week rank delta. Used to surface breakout games.

composite_rank — final ordering used by the sitemap, the home page, and the top-games prerender set. Combines popularity_score + trend_score with a freshness decay.