Moving to doing as Shay has already started working on this

We are limited to the past 90 days for useragent data that includes iOS versions, so for this upgrade distribution I used iOS version 14.7 (Release date: 2021-07-19 and a subsequent release of 14.7.1 on 2021-07-26, these are combined in most data sources). In order to compare uptake rates I used data from druid.pageviews_hourly, for edit/editors data mediawiki_private_cu_changes and used Apple App Connect data for all users/Sessions.
@nettrom_WMF made a Superset table for the druid.pageviews_hourly, data from that query and my other sources is in a GDoc Apple Relay Timeline Data iOS Version Upgrades T292104.

Biggest takeaway reflected in all 3 data sources - the period between 21 and 28 days is where upgrade distribution jumps most noticeably - from average 35% of distribution to average 66% of distribution. My completed iOS version upgrade over time matrix is in the GDoc we are using for meeting 2021-09-31, once results are discussed I will add to this ticket.

Notes on datasets: The edits/editor dataset from mediawiki_private_cu_changes is quite small (under 1500 daily) and is limited to prior to 2021-08-31. The uptake rate in that datasets is a less smooth curve than the other 2 datasets. Also note, the data we are using from Apple is Opt-in Only data, our average Opt-in rate is ~35% of all users, we are using sessions counts so we have quite enough data.

Thank you very much @SNowick_WMF!

@JMinor: Does Shay's comment above (T292104#7390329) and the "Expected Timeline" tab in https://docs.google.com/spreadsheets/d/1i-rKwQrPyDmSVFret0cKiiHetiIFuYfUWlB5tBgUr2U/edit#gid=1041396000 answer this question?