Aphorisms for Data Analysts

An aphorism is a pithy saying that reveals some astute observation or popular notion, whether true or fictitious. “Lies, damn lies, and statistics” you’ve undoubtedly heard. If you’ve taken Stats 101, you probably know that “correlation doesn’t imply causation.” Here are a few more for your consideration.

I’m being an afurism.

  1. No data analyst is an island. Be sure you have the resources you need to complete an analysis. Get help if you need it, but never pass up the opportunity to learn new things.
  2. Quality begins with staffing. Support those who generate your data. Pick good people. Give them clear, well written procedures. Train them. Provide positive feedback. And thank them when they’re done.
  3. Start at the beginning. Plan your work. Work your plan. Expect the unexpected. Ask the right questions. Beware of preconceived notions. Fairly consider alternatives. Sometimes you get lucky, just don’t count on it.
  4. Applied statistics is about compromise. The way an analysis gets done often depends as much on the administrative details as the technical details. “Stay within budget.” “Finish on time.” “Don’t collect any more samples.” “Don’t tell anyone what you’re doing.” Most mind-bending, though, is that every stakeholder will have different agendas but you won’t always be able to tell what they are. Be patient and don’t lose sight of your goal.
  5. Statistics need a good foundation. You’ll never find the answer if you don’t measure the right things on the right subjects.
  6. Precision trumps accuracy. You can’t understand the data without controlling variance. You can’t control variance without understanding the data. Variance doesn’t go away just because you ignore it.
  7. Know your data. The more you know about your data the better your analysis will be. Understanding the big picture leads to better answers but failure lurks in the details.
  8. Samples are like potato chips. You can never have just one. You always want more than you have. And by the time you think you’ve had enough, you’ve had way too many.
  9. More data are better but better data are best. Garbage in; garbage out. Get the right data and get the data right before you start the analysis. Conducting a sophisticated analysis of poor data is like painting rotted wood. It won’t hold up to even a cursory inspection.
  10. Statistics don’t replace common sense. Don’t sacrifice good sense for convenience. You can’t turn off your meat computer just because you have a few silicon chips at your disposal.
  11. Have many tools; use the best one. There is usually more than one way to analyze a dataset. Use the best tool you have or get a new one that’s appropriate. A mechanic who only uses a wrench isn’t a very good mechanic.
  12. Know when to fish and when to cut bait. If at first you don’t succeed, try and try again, just don’t be compulsive about it. Knowing when to quit is usually spelled out in your budget and schedule. It’s OK if the analysis raises more questions than it answers. That’s part of knowledge discovery.
  13. There’s a reason analysis begins with anal. Always evaluate the validity of your assumptions, your data scrubbing, and your interpretations. If you don’t, someone else will.
  14. Statistics are rarely black-and-white. Correlation doesn’t necessarily imply causation. Statistical significance doesn’t necessarily imply meaningfulness. Accuracy doesn’t necessarily imply precision. Be sure you understand what the numbers are really saying.
  15. No result is better than the way it is presented. Even the right answer doesn’t stand alone. An analysis is only as successful as the use to which it is put.
  16. You can’t always get what you want. When you don’t get the answers you and your client want, remember that there are more things to consider than just the numbers. Even a flawed study can yield valuable results.
  17. Archive your work. You never understand the important of an analysis you conduct until much later. It may be a method you used, problems you overcame, results you found, or text you wrote. Save your work in more than one way. Storage formats come and go.

You can find more aphorisms at Aphorisms Galore.

Read more about using statistics at the Stats with Cats blog. Join other fans at the Stats with Cats Facebook group and the Stats with Cats Facebook page. Order Stats with Cats: The Domesticated Guide to Statistics, Models, Graphs, and Other Breeds of Data Analysis at Wheatmark, amazon.combarnesandnoble.com, or other online booksellers.

About statswithcats

Charlie Kufs has been crunching numbers for over thirty years. He retired in 2019 and is currently working on Stats with Kittens, the prequel to Stats with Cats.
This entry was posted in Uncategorized and tagged , , , , , , , , , , , , , . Bookmark the permalink.

1 Response to Aphorisms for Data Analysts

  1. Pingback: Searching for Answers | Stats With Cats Blog

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s