Why Study Statistics?

There are lots of different ways somebody could answer this. That being the case, I'll answer it for myself. But you will likely have your own reasons.

Statistics gives you superpowers

For me, the main reason to study statistics is that it provides tools that allow me to unlock important insights about phenomena that I and my employers care about. For example, I used to work in the travel industry, and in that industry it's important to understand bookings: how to forecast them, how to detect when they've majorly dropped (like when the web site is down), and how to detect when they're "soft" (lower than usual, but not so low that it's immediately obvious. It turns out that we have to account for all kinds of things in working with bookings data:

  • cyclic behavior at different time scales (daily, weekly, annual)
  • general business trends that push the general counts up or down (e.g., shift to mobile technology, appearance of new market entrants, evolving relationships with suppliers, marketing channels and other partners, evolving regulatory landscape)
  • random fluctuation that's just inherent to real-life data sets
  • external shocks, such as COVID-19
  • missing, duplicate, conflicting and corrupt data
  • having sufficient and relevant data to draw reasonable inferences
  • and lots more...

Using statistical techniques, my collaborators and I were able to write software that's better able to convert messy data sets into a strong understanding of real-time bookings health.

Statistical skills are in high demand

In line with the above, another reason is that there's a high demand for statistics and data science skills, and such roles are well-remunerated. The U.S. Bureau of Labor Statistics released the following job growth forecasts for the years 2019-2029:

Fastest growing occupations in the US between 2019-2029, as forecast by the U.S. Bureau of Labor Statistics

Statistics is fun

OK, I admit that this one may be a stretch for some readers, but bear with me.

When we're working with data, we're working with the output of some underlying data-generating process that hiding secrets waiting to be discovered. "Data-generating process" sounds jargon-y, so let's return to my bookings example above. There, the "data-generating process" is just everything that goes into customers either booking travel or not doing it. There are so many different things in play:

  • What time of day, week, and year is it?
  • Is the web site functioning properly?
  • Are our marketing campaigns running properly?
  • How highly is Google ranking our landing pages?
  • Are we offering any deals? Are our competitors doing so?
  • Is today a major holiday?
  • Is there a major sporting event, like the World Cup or the Super Bowl, going on?
  • Is there a global pandemic going on?

You get the picture. The impact of any one of those factors on bookings is sufficiently complicated that no one can plausibly claim to know every detail. Taking them all together, it's probably safe to say that there's not a person on the planet who completely understands the mechanics at work.

Which means that studying data with statistics is largely about discovery. And discovery isn't simply profitable—it's fun! When you have a new data set, you get to explore its behavior, study and speculate on the causes of that behavior, apply statistical techniques to validate or reject your ideas, and basically learn more about what's happening underneath the hood. For me, it's highly rewarding to learn not only about my data but also about the brilliant techniques that smart people have created for working with data.