The Data Analyst in You

Everyone is a data analyst. Even you. Especially you.

We are in the midst of an era where terms like "big data", "analytics", and "data science" are tossed into our professional lingo as liberally as the amount of Parmesan I toss onto my pasta. In the past decade, data's role within many organizations has evolved from a luxury to a necessity. It may feel like the only thing worse than not being confident in your math skills is publicly revealing this lack of confidence.

You are not alone. 50% of SAT test takers scores below a 500 on the math section, and a 2005 Ipsos poll revealed that 4 in 10 adults said that math was their most hated subject in school.

One of my roles as a statistician is to make sure all questions are answered and all questions are asked. That means we are a team. My expectation is that you will have a role in my analysis, whether its a study of the past or a prediction of the future. If you, my teammate, are not understanding what I am presenting, we are both held accountable.

Here are some useful tips I always share with my number-loathing colleagues to make them into good data analysts:

Don't Get Caught up in the Numbers: Here is a little secret: mathematics and statistics are different. The mathematics are important in statistics, but not to you. Don't be overwhelmed by the numbers that might bring up flashbacks of hours spent scribbling on loose leaf paper in junior high. What is important for you are the relationships: what do the numbers mean for you? What information went into calculating those numbers, and what should you do with them now that you have them?

Context is King in Cause and Effect: You are the subject matter expert. The data scientist/data analyst/statistician is looking at the data relationships and does not have your experience. Provide context. Context can help the statistician look at new perspectives or create hypothesis to test. Most data analyses are answering some form of cause-and-effect question, but, more often than not, the data is only a proxy or placeholder of the true underlying cause. That means context is needed to understand data's story.

Don't Get Caught up in the Data: Do not let your knowledge of the data constrain the questions you ask. You have many questions, so ask them. It is our job as a team to figure out how to answer the questions. The solution might not be obvious, but if you have a good data person, that person should be creative in manipulating existing data in a new way that, at the very least, might indirectly answer your question. If there is no way to answer with the data available, that means you can figure out how to gather the necessary data for the future.

Don't Leave Without Confidence: Look again at the data I presented in my third paragraph: people do not like numbers. You are not the only one, but you are the one who might have to act on the results. Get the clarification you need to be as confident as the statistician. If you feel uncertain, that means we have work to do. It is my job as a statistician to make sure you are using my analyses correctly (my reputation is on the line too), and I do not consider my job finished until I answer all questions and all questions are asked.