Meanwhile, while HN bubble imagines people doing big data jobs on Excel, in the real world 10s or 100s of millions of people are perfectly satisfied doing small data jobs in Excel.
I know i have had to correct one in three excel sheet i have ever gone over using pen and paper in order to validate the results but i am a paranoid sod who actually do this kind of exercise on a regular basis.
almost all of the disciplines known to rely on excel have a serous issue with repeatability of results either because nobody ever attempts it, or because it's a messy field without a well defined methodology.