The title of our section is "Section for Statistical Programmers and Analysts." As individuals, it's not really an either/or. We are both programmers AND analysts.
Recently, I was tasked with comparing two datasets and delivering a report on discrepancies. These datasets were created by two different groups with two different approaches. The datasets varied in both structure and in content, although they shared a common design target.
I started to use typical "compare" procedures and was getting nicely formatted reports. Unfortunately, they were really difficult to understand. Individual pieces were clear enough, but the big picture (if any) was not apparent. It was hard for me to get to the "gestalt" of what the discrepancies were.
After struggling a while and refining my program several times, I found myself unable to clearly explain what I was seeing. After some reflection, I stopped programming and started analyzing.
I chose a few representative clusters of records (in this case, tied to clinical trial subjects) and went through a manual comparison between the two deliveries. In this manner, I was able to make sense of the discrepancies and provide a text report which was useful to my internal client.
Programming took me only so far in this project; I needed my analytical skills to truly finish the job.
As we work on projects that come our way, I think it's important to look for opportunities when programming can stop and analysis can begin.