I have performed logistic regression in Calc, by using the Solver feature to find coefficients, after multiplying the input with them, passing the output to a sigmoid, and computing the mean squared error.
I suspect anything more advanced than this would be better worth implementing as a script, rather than using Excel.
The main arguments for Excel, that I suspect, would be better collaboration features (like comments and sharing and such), and better integration with other MS tools (Outlook etc.).
But even then, perhaps a script or Jupyter Notebook would be easier to work with (for diffs and version control that are not vendor-locked).
>
The main arguments for Excel, that I suspect, would be better collaboration features (like comments and sharing and such), and better integration with other MS tools (Outlook etc.).
No, the main argument why Excel is used in the finance and insurance industries is that many users know Excel really well.
I suspect anything more advanced than this would be better worth implementing as a script, rather than using Excel.
The main arguments for Excel, that I suspect, would be better collaboration features (like comments and sharing and such), and better integration with other MS tools (Outlook etc.).
But even then, perhaps a script or Jupyter Notebook would be easier to work with (for diffs and version control that are not vendor-locked).