The One Trait All Successful Data Analysts Must Possess

When people inquire about my career path, the words that most frequently tumble off my tongue are “unusual” and “fueled by curiosity.” Here’s what I mean.

My undergraduate degree is in Business Management, and I quite literally fell into my first job as the accounting manager at a community hospital (even though I knew absolutely nothing — zip, zero, zilch — about medicine, hospitals, or anything to do with healthcare). Additionally, even though I had earned passing grades in my accounting courses, I hadn’t loved them (no one in their right mind does).

Fast forward to my decision (some might call it a masochistic whim) to pursue a graduate degree in clinical outcomes and quality improvement. Unlike just about every other student in the program, I had absolutely no clinical or research experience. Nevertheless, I was accepted, and completed my degree. Along the way, I also became deeply interested in data visualization. As there was (there still is) very little formal training in that field, I simply read and researched and learned everything possible from everyone I could. I just couldn’t help myself.

At each juncture in my professional development, I was not the perfect candidate on paper. I didn’t have the exact credentials or experience required. But what I did bring to each stage along the way was a deep, lively curiosity, and that has made all the difference.

So what is it about curiosity that makes it such an important trait — especially in healthcare data analysis — and how can you identify it in job candidates ? Quite simply, curiosity is being inquisitive and interested in something– the urge to ask why, and the unwillingness to settle for the first, the brief, the (usually) superficial answer. When we are curious, we question, probe, and inquire — all actions that are crucial to uncovering interesting and important stories in our healthcare data. If the data analysts we hire aren’t blessed with this intense, even relentless, curiosity about their work, it shows: the reports we receive lie dead on the page, defended by comments like “I gave them exactly what they asked for” (translation: and nothing more). Even worse, the data may not reconcile; the results make little sense and provide no insight.

The challenge, of course, is to be able to figure out before the hire who is seriously curious. Traditional interviewing techniques won’t cut it; you’ll need to add something new to the mix.

Here is one method that we’ve found extraordinarily helpful: the case-study interview. This type of interview is a two-way conversation that demonstrates a candidate’s ability to think and solve problems creatively. It is truly useful in identifying the candidate who sees solving a case or problem as intellectual stimulation and challenge — as fun. Such a person is very likely to be (you guessed it) seriously curious.

Here are the steps we use in designing and conducting this type of case-study interview:

  • Create a simple spreadsheet (we use MS Excel) of some of the data elements you typically work with (stripped of identifying features, but accompanied by brief data definitions). Add a short case study with four to five questions you would like the candidate to answer. (Build a little ambiguity into the questions: you want this to be a bit challenging, not rote). Present it to those candidates you have decided to interview, allowing each the same amount of time to prepare an analysis and presentation. Be sure to specify the amount of time they have to present and whether you want them to leave hard copies with you.
  • Have each candidate deliver a 15- to 20-minute presentation to the group who will be making the hiring decision. This part shouldn’t be wildly complicated: it isn’t about software applications or statistical methods. Here you need to focus on how a potential candidate approaches the case; presents the data; and discusses what (s)he’s discovered-and what wasn’t well understood.
  • Additionally, you or a member of your current team should prepare a summary document of how you would have approached the same data and what you would define as evident (the basics) as well as what is potentially interesting. This will serve as a comparison and guide for interviewers who may receive reports, but who are not experts in how data can be analyzed.
  • Aim to schedule a number of presentations on the same day, so you can discuss and compare presentations while they’re fresh. Ask each interviewer to score candidates on how well each answered individual sections of the case, then compare and discuss your results.
  • Prepare to be amazed.

We have worked with several clients to develop and conduct these case-study interviews, and without exception they are always astonished at what they hear and see and learn about candidates. They are amazed at how easy it becomes to tell who has the curiosity trait, and which candidate went after the data with real interest.

To be successful at anything — and most assuredly at healthcare data analysis — you must be curious. When you never tire of saying, “why?” you remain interested and engaged for the sheer fun and excitement of finding an answer, discovering new information, identifying fresh opportunities, and coming up with innovative solutions.

Perhaps Walt Disney said it best: “We keep moving forward, opening new doors, and doing new things, because we’re curious and curiosity keeps leading us down new paths.”

Where will curiosity lead you — and your colleagues — next?

This Post Has 2 Comments

  1. Bill Droogendyk

    Kathy, I think that you are right on with the curiosity theme.
    As I work with people to help them create better DV displays, I’m finding that one key to good DV is having that curiosity to find the story in the data. If the story isn’t understood, we’ll never be able to tell it. At the same time though, being able to sketch out various ways of looking at the data does help to reveal the story – and curiosity will drive us to look at the various ways – and reveal the best way to tell the story.
    Keep up the good work!

  2. Bill Droogendyk

    I would also add this to “translation: and nothing more”: “and not what they needed to tell the story that’s in the data”

Leave a Reply