R Learning Curve for an Excel User

What Are Fellow Data Visualization Bloggers Saying About R?

Here’s a quick rundown on what three fellow data visualization bloggers have to say about the R recently:
 

Tony Rose at Support Analytics wrote this comment on my  Data Loss Aversion post “..I agree that Excel was not designed well for advanced charting and have been meaning to dive more into R.” 
 

Derek at Information Ocean  followed up Tony’s comment with  ”I too would like to get started but haven’t been able to crack the starting barrier yet. “
 

Robert Kosara at EagerEyes wrote an interesting post on the Electoral College and popular vote outcomes. He showed an interesting Excel bar chart that took quite a bit of time.   Hadley Wickham, the developer of ggplot2, commented to Robert’s post that “Maybe you should learn R “. Robert replied “..You’re right, I really need to get around to doing that. I have it sitting on my computer, but just haven’t gotten around to doing anything with it.”
 

Derek replied to Robert’s comments with “That is exactly the situation I’m in at the moment. I have R installed, and now I’m wondering when I will ever have time and thinking-space free to break through the learning barrier.”
 

Posts and comments like these tell me that there are others like myself who want to move up to R and are finding the learning curve challenging. Having spent several years mastering Excel charts and VBA, I see the R  learning curve as similar to what I experienced learning to make effective charts with Excel and VBA.

The only difference is that with R, I have a much more powerful tool with 100’s of world class programmers developing new tools all the time.

R has many multivariate chart and analysis capabilities not included in Excel:

• Dot plots
• Step charts
• Box & Whisker plots
• Trellis – Lattice – Small Multiples 
• Lowess smooothing

I have been able to develop rudimentary versions of the 4 advanced chart types, however, I have not been able to prepare a lowess smoothing program in Excel – VBA.  Once I started comparing R to my Excel VBA tools, I realized that I could produce a better, more advanced chart with one line of R code than with dozens of lines of VBA code. By switching to R for my multivariate charts,  I gain from the work of 100’s of world class statistical programmers, letting me concentrate on producing advanced charts rather than developing Excel VBA code that just handles part of what the R function does.

R Learning Strategy

Having tried to learn R several time before, I’ve developed an R learning strategy that I think will work this time:

• Use Excel charts for all my simple, straight forward charts.
• Use Excel for all of my data manipulation. Since I can do just about everything I need in Excel, why relearn data manipulation in R
• Use a friendly R interface tool. I selected TINN-R, a simple and effective tool for writing my scripts.
• Use the free, Chapters 9 & 10 from Introduction to Data Technologies  written by one of the leading statistical charting programmers, Paul Murrell. 
• Transfer data to R through CSV files. It’s easy and it reduces my learning curve. 
• Learn just one of R graphics packages, in my case, Lattice. It does a great job on trellis – lattice, multivariate charts, Excel’s Achilles heal.
• Use a code snippet tool (VBCC in my case) to store my R code snippets in an easily retrievable form. I used this same tool to save my VBA code snippets, saving me lots of time on my VBA learning curve.
• Keep going! Now that I have a few scripts under my belt, I feel much more comfortable.