We fetch the App Store Analytics – App Installs & Deletions (Daily) report via the Analytics Reports API and land each delivery into S3. The rows we ingest contain fields like Date, Event, Download Type, Counts, and Unique Devices. We’re trying to compute for example Installations and First-time downloads so our warehouse totals match the App Store Connect UI exactly, but our counts are consistently higher when we aggregate (with filter conditions being the same).
Questions:
(1) Official dedupe/versioning approach: What's the recommended approach to avoid double counting across API deliveries and to match the UI exactly? If later deliveries can revise past dates, what field(s) or process should we rely on to decide which records to keep for a given Date?
(2) Completeness window: For daily data, is the recommended approach to publish rolling values that may change for a few days, or waiting until dates are considered “complete” before reporting? Any best-practice guidance would be helpful.
Thanks!