Looking at Steem-apr more closely it seems like it suffers from the same miscalculations as SteemDesk. SteemDesk takes all payments into account, even those which do not come in as a return for the initial investment. This calculation results in wrong APR reports.
This is something I wanted to fix in SteemDesk for a long time. Unfortunately, I currently do not have the time to work on SteemDesk.
Agree, the transactions should be filtered separately but it is difficult job (template for each service is needed). I believe our simplification is enough to estimate revenue.
Yes, it is a lot of work, but it would be useful to have filters that could be applied by the user for example.
I'm gonna need to check if @wefund or @lavater receive any transfers from the services that are not payouts. Maybe I can help here somehow.