The POP (plan operator level) is as deep down as it gets for the non-core HANA developer.
What any POP is supposed to be doing can typically be found out by hovering over the box and checking the pop-up (pun intended) information.
With the information available it's not possible to say for sure what's going wrong here.
However, looking at the high amount of memory used and the large number of processed rows, my guess would be that there's a lot of data being moved (and eventually being materialised).
That's something that should be avoided for the obvious reasons.
Based on your description I doubt that your model is designed in a wise way - union the output of calc. views (likely aggregating calc views) can easily lead to materialisation and the BS* tables really invite to have a too narrow grouping condition resulting in too many groups (= output rows).
That's as much as I can say based on the available information.