10 ms·
For many uses, I dropped Excel and replaced it with R. Doesn't hang with thousands/millions of rows.
by partycoder 10y ago
For many uses, I dropped Excel and replaced it with R. Doesn't hang with thousands/millions of rows.
- roel_v 10y agoOnce you know what to do, yes. But for scenario analysis and prototyping, anything with a code/run cycle is just painful. I've tried switching many times. Plus, Excel has a 'gui' (of sorts) build in. (Shiny is not a substitute for that - here too, shiny is fine for productizing once you know what you want to do, but not for quick one offs)
- em500 10y ago> for scenario analysis and prototyping, anything with a code/run cycle is just painful For scenario analysis and prototyping, you'd use RStudio as GUI and the command line for REPL. But personally, I find Jupyter notebooks with Pandas even better.
- roel_v 10y agoRStudio is not a gui, it's an ide. If you change a parameter somewhere you still have to re-run the code and hunt for whatever it is you're calibrating. In Excel, you just change values and you see immediately what has changed. In Excel, you can trivially show subsets (with a dropdown that show you all values in a range), or checkboxes for booleans, do conditional formatting to highlight special values (< 0, within ranges etc). All of which is a pita with R and with anything that has a code/view report cycle, basically.
- em500 10y agoThe table viewer in RStudio shows any changes in dataframes live as you change them in the REPL, so I don't really see your point.