If you have Baloo running on your system already and it has indexed _most_ of the files, it would make much sense in my opinion to

1) use Baloo's index to gather files / the sum of their sizes 

and

2) for files and folders Baloo hasn't indexed yet (most likely because they were excluded from the index), fallback to the original recursive search + as a bonus update Baloo's filename / xattr index in the same step (excluding excluded folders)

This way generating the charts could become way faster for larger file systems, especially on slow devices (HDDs for instance) and you'd not need to build up two separate indexes, which largely overlap.