Regarding your first question: I would say that this is a practical example of geophone (weak motion detection) clipping vs accelerometer (strong motion detection) non-clipping: https://twitter.com/GeoGinger/status/1426580715416825861/photo/1
Clipping happens a bit above/below ±0.02 M/S for a geophone. You can explore this by going here Data View: Raspberry Shake Data Visualization Tool and selecting August 14th, 2021 as a date.
The event at ~12:29 UTC is clearly clipped on the EHZ
(geophone) channel, but is not on the EN*
(accelerometers) channels.
For your question #2: The arrow is required as our 3D and 4D models have to be oriented with the arrow pointing towards the North, so that the horizontal geophones/accelerometers (North-South and East-West) can detect waves in the correct orthogonal orientation. So, as you correctly said, ENN will be along the NS axis, and ENE will be along the perpendicular EW axis.
There are a lot of threads on the forum about general knowledge: Notes on how I use my Raspberry-shake but I also encourage you to check our Educational resources: https://edu.raspberryshake.org/ and the pages from the old IRIS (now EarthScope Consortium): Education and Public Outreach | IRIS