Nothing that’s not easy to fix 
The first issue is that it looks like you’ve been averaging percentages, which doesn’t work right. You actually need to sum up all the counts and divide by the total number of opportunities to get the correct percentage.
The second issue is that the reported percentages on titans should be given per loot type slot, rather than per titan. That way, the result is fully portable to higher or lower level titans.
I quickly ran the numbers on the data you have (thank you so much for making the underlying data available!). Here’s what I get on the AM:
There were 265 total reported AM in the data set. Within that, the breakdown is:
12 4*
58 3*
176 2*
19 1*
That means if we divide through by slots, we get:
4* AM: 4.5% per slot
3* AM: 21.9% per slot
2* AM: 66.4% per slot
1* AM: 7.2% per slot
This is actually very much in line with the 4* AM data that @kamikaze_assassin reported for 12* titans from CP.
You can likewise total up the number of ingredients, battle items, etc., to get the number of loot slots of each kind you all saw, then divide by those to get the number of 4* ingredients per slot, 3* ingredients per slot, etc.
ETA: if you’d be willing to share the sheet with me, I’d be happy to make a copy of it and tack in the changes I’m suggesting in the copy.