My start with R

The programming language that made me a better scientist

Mina Kmiecik, Deep Blue, 2023, acrylic on canvas.

After I was finished with my undergraduate degree, I took a year to gain more research experience and prepare my graduate school applications. I was working as a laboratory manager and occasionally helped graduate students with their projects.

One of these students showed me his data analyses that were performed using a statistical programming language called R. He showed me the code that he used to wrangle his data, perform statistical tests, and generate figures. I remember seeing white characters on a pitch black screen. It was like something from the movies. A hacker breaking into the main frame. It was completely foreign to me. 

He scrolled through a folder that had PDFs of the figures he created using R. He showed me a few of these images; they looked clunky and rigid. The numbers from his analysis were in one place, while the graphs, each in a separate document, were in another. 

I thought it was completely unnecessary and made things more difficult than they needed to be. Why wasn’t everything integrated into one output? Why would someone do it this way?

“Why don’t you just use SPSS?” I questioned him. (FYI SPSS is a software program developed by IBM. Think an expensive and fancy version of Microsoft Excel). I figured that programming was not necessary anymore these days (this was over a decade ago!) and we had software that made things easier on us as scientists. No need to learn esoteric languages; we have SPSS. Just point and click! 

He was kind to me and patiently explained the advantages of R. It was an open source language that was free to use. But his explanations were lost on me. I couldn’t see the bigger picture. I was not open to learning new things.

Looking back on this memory, I can only shake my head at my naiveté. It’s been 11 years since this first experience with R, and now I can’t imagine my life without this wonderful programming language. It’s become one of my favorite tools in science. My technical skills in R have presented me with jobs and opportunities ever since I decided to teach myself the language. R also helped me learn statistics at a deeper level—an entirely unintended consequence. 

My complete rejection of R at first may surprise some that know me. I was a huge proponent of the language in graduate school and into my first postdoc, trying to teach anyone and everyone who would listen. It’s one of the main reasons why I started a website/blog in graduate school (built entirely in R of course).

I entered graduate school with experience using SPSS to analyze data. I did most of my data visualization using Excel, and I thought that the SPSS + Excel combo was just the way things were done in my field. I received more formal training in SPSS my first year of graduate school and I was flying high, deep in analyses for my first project. 

Then, things started to take a turn for the worse. 

Financially, grad school was tough and it was time to pay for another yearly license of SPSS. I don’t remember the price for a year subscription, but it was enough for me to emit an audible gulp when I visited the website and saw the renewal fee. Another obstacle was that scripting in SPSS was becoming burdensome. The script for my analyses was over 1000 lines long and it crashed 80% of the time. It also took a very long time to run on my machine. Every tiny update I made cost me another 20 minutes; I needed answers fast!

A thought began gnawing at me. What about R? It’s free and fast, but I knew it was going to take some time for me to learn. Our statistics professor showed us a few things using R, so I had somewhat of a background on getting started, but I wasn’t sure whether the time I would need to take away from my research projects would be worth it in the long run.

One day when I was working in the lab, I was under pressure to get an abstract submitted to a conference and I had only 24 hours to go. I needed to make some updates to my analyses, run the SPSS script, examine the SPSS output, and then write the abstract with the new numbers. Simple enough.

I ran my SPSS script and went to make myself a coffee. But by the time I got back to my desk, the script had crashed. “No problem,” I thought as I clicked the button to run the script again.

This time the script crashed immediately. Something was wrong. I spent the entire day trying to get this SPSS script to run. I managed to squeak out an abstract with the updated numbers, and have my advisor review the changes, just in the nick of time. Only ten minutes to spare before the online submission portal closed. Phew!

The next day, I quit SPSS cold turkey and started learning R. I even uninstalled SPSS from my machine to quell any attempts from using the software for my work. The decision wasn’t easy though. I knew that learning R would take several months and this would stall the progress I had made on my research projects. All things came to a grinding halt.

The transition was painful. The learning curve steep. Every time I hit a wall with the most simple data analysis task I felt the urge to abandon R and go back to my comfort zone of SPSS and Excel. But then I thought about the crashing SPSS scripts and paying for renewal licenses with money I didn’t have.

But I persevered and came through on the other side a better scientist. I may write about the process of learning R and what that was like for me at a later date; however, I’d like to make a slightly different point here. It’s a lesson that I learned during my first foray into R; a lesson that I seemed to learn over and over again since then.

Have the courage to step out of your comfort zone to learn a new process that may improve the way you currently do things. I have found myself at this fork in the road several times. To the left I see a sign reading, “Stay on the current path and the work will be easy and get done faster.” To the right the sign reads, “Learn this new skill that will be difficult and stop your progress.”

Every time I’ve taken this slower and more difficult path to the right, I’ve never regretted it. The path to the left is comfortable and you get things done quickly, but then you come out the same as you were before. The path to the right pays future dividends that are unimaginable and invisible in the confines of the comfort zone. 

It’s easy to get wrapped up in the pursuit of progress and completing tasks. But the next time you find yourself at the crossroads of comfort, try to trade off progress for growth. Your future self will thank you.