Microsoft Excel has long been used by both managers and engineers for quick data analysis. Entire dashboards have been built with this tool and discussed daily in production meetings. The convenience of scrolling through the data and the ease of making charts without using any code makes users reluctant to part with it, even though better alternatives exist.

I’m not against using Excel. I’m against using Excel as a database: copy-pasting data and emailing this data around. As long as Excel reads from one consistent source (database, data lake, etc.) becomes just another viable tool to analyze data.

An additional problem is that calculations are not explicit: we just see the graph and not the data manipulations that were made beforehand. So we have no choice but to trust (or distrust) the author. This issue of reproducibility is a separate discussion.

In November, an interesting Delta plugin has been released: it connects Excel directly to your data lake and reads data from Delta tables into an Excel worksheet. Cool stuff. If your company uses Databricks and you prefer Excel over coding in notebooks, I suggest you check it out.