Mac App Store Apps Dataset Data: Dataset Structure and Field Coverage

Apple's terms also govern how the returned assets appear in products: app icons and album art may be shown only adjacent to an approved badge linking back to the purchase page, previews must be streamed rather than downloaded, and content must be removed on Apple's request. The API is documented under Apple's archive pages, governed by California law, and remains free of authentication at every tier that matters for a monitoring job.

The 45 columns fall into five working groups, and each supports a different analysis:

1. Identity: app title, seller name and genre, which is what you group by when sizing categories.
2. Commerce: price fields that make storefront-level pricing arbitrage visible across the US, German and Ukrainian listings.
3. Reputation: rating values and counts collected at crawl time.
4. Footprint: binary size and supported device lists.
5. Assets: artwork URLs, handy for building visual catalogs or training image models.

Desktop numbers land differently when set against the mobile catalogs in the same industry:

The comparison exposes a structural gap: the appGoblin table carries installs, ad_supported and in_app_purchases flags across its 15 columns, while the macpaw-research table has no install counts because Apple's storefront does not publish them. Any vendor-size claim about the Mac catalog therefore leans on row counts (61,196 listed apps) rather than adoption figures. Note also that appGoblin's Apple slice spans roughly 1.50 million iOS apps, an order of magnitude above the desktop catalog.

Four workflows fit this data well. Competitive mapping works out of the box: group 61,196 rows by seller and genre to see which vendors maintain multiple Mac App Store titles and where categories thin out. Pricing research uses the three-storefront structure to compare identical apps' US, DE and UA prices side by side. Portfolio and market-sizing studies join the static table to live /lookup calls, flagging price changes, rating drift and new versions since January 2024. Asset-driven projects - icon galleries, thumbnail models - draw on the artwork URL columns directly.

For context on whether the platform deserves the effort, StatCounter Global Stats publishes free usage-share time series for operating systems and mobile-versus-desktop splits from January 2009 onward, exported as CSV per stat under commercial delivery terms-SA 3.0. It measures usage rather than catalog composition, which makes it the denominator-side companion to the numerator-side app table. One caveat worth stating plainly: because neither source tracks installs on macOS, revenue estimates require assumptions the data itself cannot supply.