I never though that writing SAS or SQL code would compare to writing a paper. I've been deep in developing a new report during the past few weeks. The similarities have been striking.
When I wrote papers in college, usually the assignment was quite vague. Something along the lines of "pick a topic about X and write on it." Choosing the right topic was half the battle. Specific yet with enough room to play with my own thoughts on the matter. Something people had considered, but not overly considered.
Then I would be off to the library like a pioneer in search of a new road through the mountains. Exploring where others had already tread, trying to find my own way at the same time. After hours of reading and note taking, I would wrestle with my own opinions about the topic. Then, after sleeping on it for a few days where my subconscious was exploring different theories and approaches, I would sit down and give birth to a paper.
It was a thrilling process for me. Usually I love every minute of it, but especially the research and the percolating phases.
I've come to see that writing analytical reports is not that different. The client has a vague idea of what they want. A general question that should be answered. They don't know how to best answer it though. It's up to me to see how similar questions have been answered before. To wrestle with the nuances of this particular problem. To find ways of answering the question that the client has never even thought of.
Then it's time to write the code. Usually the first round of code-writing gets at 90% of what is needed. The last 10% requires the most work though. It's where I become a detective, trying to think of how many ways can I break the report. How many irregularities can I tease out of the data? What do those irregularities mean for the client? In many ways, this detective process is the same as proactively combating opposing arguments to the thesis of your paper.
Both in paper writing and report coding, I am the creator and the artist.