Probably something like this

https://youtu.be/Ynvoriv09Ks?t=105

Ah, is this Ryanair's plan to divest from their all-Boeing fleet? :)

It should be pretty simple for Flightradar24 to exclude non-commercial aircraft from the data through, which would solve that problem.

There's also tons of data available in the ADS-B signal that should help distinguish between aircraft-motion-induced outages and actual jamming: https://mode-s.org/decode/content/ads-b/7-uncertainty.html

From the FAQ, it sounds like they simply presume that anywhere with multiple low NIC values is indicative of interference.

> The GPS interference data is derived from NIC (navigation integrity category) values that we receive as part of the ADS-B protocol. We mark regions as affected if a significant number of flights in that area report lowered NIC values.

Unfortunately some of the data that's most directly applicable to determining aircraft attitude, like roll, is optional and rarely sent by aircraft, but yes I'm sure you could do a decent job of inferring maneuvering from change in heading and vertical rate (especially if you're looking at ADS-B data with high temporal resolution vs., say, every 10-60 seconds.)

on a pure radar scan, what would return of formation flying like this look like to a radar operator? is it just one large dot, or can they distinguish the number of planes in the formation?

It happens when you are giving the bird.