About that scaling…

I’ve seen workflows with data chunked out into multiple excel files because the results exceeded the row limit (4 million?) Going through all that by hand was a pain, and it took hours to go from a flat file format to spreadsheets. Parquet took this down to seconds, and provided a rich data analysis API.

Spreadsheets scale for small to medium-ish workflows and are very much a drag for anything remotely “big.”