Believe it or not, Excel is still my go-to analysis tool a lot of times, because it’s great at what it does. I’m a shortcut fiend, so I can do things pretty quickly. So when do I opt for R? People have asked me this many times. Here is my unofficial checklist I loop through in my head to decide whether to Excel or not to Excel:
- Is the data not well structured or PivotTable-ready? Does it have a lot of stuff within cells that needs to get broken out?
If yes, then R, unless I can work my Excel magic to clean it up.
- Is this a quick and dirty one-time analysis? Including quick visuals.
If Yes, then Excel, as long as the data is not gigantic.
- Do I need anything beyond basic statistical analysis? Regression, clustering, text mining, time series analysis, etc
If Yes, then R. No contest.
- Do I have to crunch a few disparate datasets to do my work?
Depends on complexity. If data sets are small and a simple vlookup can handle it, then Excel. If more than three tables, most likely R. If more than 1-2 columns vlookup’ing from each table, also R.
- Something I will want to share in a web-based, interactive format that is nice to look at?
R with the Shiny framework
- Unique and beautiful visuals the world has rarely seen?
While I was learning R, I used a hybrid approach … doing the heavy-lifting data prep work in R, then using the write.csv() function to send my data frames back to Excel for visuals and basic analysis. Over time, I have learned to do more complete analysis in R, from beginning to end.
I hope this helps! What scenarios did I miss?