Feature Request: downloadable data

[jamescridland]

Inspired by Podcast Index's downloadable database...

I'm wondering whether it's possible to perhaps produce a (monthly? weekly?) downloadable SQLlite dump for research purposes, based on the "downloads" data (i.e. these verified downloads, rather than individual requests).

This would enable researchers into podcasting to, for example...

SELECT agentName,COUNT(*) AS totaldownloads WHERE devicetype="android" GROUP BY agentName ORDER BY totaldownloads DESC

• see the most popular Android apps by total downloads

SELECT agentName,COUNT(*) AS totaldownloads WHERE countryCode="DE" GROUP BY agentName ORDER BY totaldownloads DESC

• see the most popular apps in Germany

... and many other permutations we've not yet considered, like normalising US downloads by timezone, or looking at auto downloads vs user-initiated downloads.

I'm imagining that the files would be, um, quite big. Podcast Index doesn't appear to store theirs, though; so you might be able to simply host "the latest one". I'm thinking that monthly might be best, but I've no idea whether that makes it a stupidly large file.

For filesize, perhaps you don't need the URL (you could look that up); and perhaps also you could lose the continentCode (given that you can reproduce that based on the countryCode).

    {
      "time": "2023-02-01T00:01:11.872Z",
      "url": "https://op3.dev/e/https://cdn.changelog.com/uploads/podcast/news-2023-01-30/the-changelog-news-2023-01-30.mp3",
      "audienceId": "3d9d84cd056ba3d4785e3d460a156a63bbf68702f8a477339951520ab6bc9420",
      "showUuid": "a18389b8a52d4112a782b32f40f73df6",
      "episodeId": "381f2d07178b6affe8acad5266b63fe847b2c5a1c919251d7dd10ba6d5eb1092",
      "hashedIpAddress": "b5ea00bdd3844c026cbfeb3dcec58f06bd94aa90",
      "agentType": "app",
      "agentName": "Apple Podcasts",
      "deviceType": "mobile",
      "deviceName": "Apple iPhone",
      "countryCode": "US",
      "continentCode": "NA",
      "regionCode": "WA",
      "regionName": "Washington",
      "timezone": "America/Los_Angeles",
      "metroCode": "819"
    }