There are a variety of setting that need to be disabled to gather good stats against the complete source systems. It would be helpful to have a way to ignore the EXCLUDE_FROM_WORLD and COMPATIBLE_* items. Ideally a system similar to the -s option to gather all the recipes, but then parse the metadata.
This is a placeholder for ongoing improvements to stats.
At this point we can gather alot of the stats that we need from the packages.yoctoproject.org site. We will continue to enhance that code, new bugs should be filed against the site.