In this notebook I show some of the systematic errors in the fundamental data, which make its use much more complicated (for me at least). I am very curious whether other users have written code to correct these errors or how you cope with it in general.
At first sight, it would seem that these errors are not present in the actual data at Morningstar, but rather originate somewhere in the transfer of the data to Quantopian, so that it should be possible (an desirable!) to correct these errors at the source. I cannot know on which side the problem lies though, whether at Morningstar or Quantopian itself.