Speaker Author Trainer Professor Advisor



The Wild Wild West of Analytics Programs

Pedagogy and definitions around higher ed's trendy new discipline are anything but consistent.

Estimated reading ⏰: 6 min

Introduction

As I write these words, I’m in the midst of teaching my fourth year of analytics courses at ASU. To be sure, it feels like longer than that. That’s probably because, during this time, I have done more than merely fulfill my 4/4 teaching load. I wrote a book on analytics, taught courses over each summer, and recorded a 7-week online analytics capstone course back in 2017. I’ve also discussed the topic with my peers both at ASU and at other universities. Finally, I’ve spoken about the topic at a few conferences in Phoenix.1

To call the last three-plus years instructive would be the acme of understatement. (Pun intended.) In that time, I’ve learned a great deal about teaching analytics. To wit, I have reached an unequivocal conclusion: Much like in the corporate world, universities that currently offer degrees and courses in analytics are still trying to figure it out.

Let me explain.

Academia plays catchup

Let’s be clear: Analytics as a practice didn’t just arrive yesterday. Individuals and organizations have been making data-based decisions for decades. Behemoths such as Amazon, Facebook, Google, Netflix, and Twitter have taken this to another level, often with disastrous consequences. Brass tacks: data and analytics are double-edged swords.

In academic circles, however, the discipline of analytics remains a relatively new and immature topic. This should surprise exactly no one. By way of comparison, consider the early to mid-1990s. It’s not like higher ed immediately nailed nascent topics such as enterprise systems and the World Wide Web. I’m sure that some professor somewhere taught a Kozmo.com case study. To be fair, we can only fully understand and teach many things in hindsight. As Søren Kierkegaard said, “Life can only be understood backwards, but it must be lived forwards.”

Analytics means vastly different things to different people

Against this backdrop, as I look across academia, two things are obvious. First, analytics matter more than ever. Exhibit A: Johns Hopkins is now wisely infusing technical courses in their MBA program. Second, there’s anything but universal agreement on the very term analytics.

Allow me to offer my own two cents here. At a high level, I’ve always thought of analytics simply as the process by which organizations make decisions and gain insights based upon data. That’s it. Beyond that, there’s a general consensus that analytics falls into one of the three buckets: descriptive, predictive, and prescriptive.

Opinions and definitions on analytics vary—often widely. Oddly, many schools offer degrees data analytics—as if there’s another kind. Cue Jack Nicholson quote from A Few Good Men.

This begs the question: Can schools, departments, and faculty effectively and consistently teach a subject if they can’t really define it?

By way of contrast, most faculty teach entry-level economics and accounting in virtually identical fashions. In these courses, professors speak a lingua franca. Sure, the textbooks and examples may vary from school to school, but all students can expect to learn the rudiments of supply and demand and debits and credits. This holds true whether a student attends Harvard or Tallahassee Community College.

There’s anything but universal agreement on the very term analytics.

Let’s get back to analytics. In meetings and when I interview prospective faculty members, I frequently hear other categories of analytics bandied about beyond the three mentioned above. Yes, I’m talking about diagnostic, exploratory, and edge analytics. (This reminds me of the “v explosion” that accompanied Big Data a few years ago.)

Where do you put analytics, anyway?

This isn’t just an existential question. The answer portends many consequences for the courses that recommend and require that students take. Related questions include:

  • Is analytics a separate discipline or is it at the nexus of data, management, and technology?
  • Should it exist in its own department or should it live “under” information systems (IS), operations, management/business, or statistics?

I’ve looked at how some schools attempt answer these questions one thing is clear: There’s anything but unanimity.

What’s the relationship between analytics and data science?

Again, this is a core question to which there’s currently no correct or even consensus answer. Adding to the ambiguity is the definition of data scientist. Again, opinions vary, but I’m fond of Josh Willis’ simple yet effective tweet.

Where does analytics end and data science begin? What courses are appropriate or imperative for each?

Again, these aren’t just theoretical or rhetorical queries. The consequences of each school’s answer drive the courses that they offer and require students to take. For instance, consider two schools: ABC and XYZ. The analytics program at ABC emphasizes technology more than business. As a result, it teaches students powerful languages such as Python and R.

By contrast, XYZ emphasizes the business side. As such, it offers an entire class on Tableau or another best-of-breed dataviz application. In this case, the teaching of higher-level programming languages becomes optional. The focus here is the “business” side of analytics, not the quant side.

What do we teach undergrads vs. grads?

Again, I see more questions than answers here. Lines at present appear very blurry and I haven’t yet heard a compelling distinction.

How many universities routinely use analytics to make decisions?

Based on my conversations with peers, many universities continue to struggle with the basics. I’m talking here about wrangling data, data quality, and governance. By and large, few universities have “graduated” (pun intended) into using proper analytics themselves. This is not lost on me. Perhaps this stems from their own decision-making processes. In other words, consider the following questions:

  • Do administrators make data-based decisions? If so, then how often? If not, then why not?
  • Are administrators willing to go where the data takes them—even if it means questioning core beliefs? (For instance and as I write in Analytics: The Agile Way, Google ditched using grade point average [GPA] when evaluating software engineers because it didn’t correlate with long-term performance.)
  • Do administrators and IT folks refuse to make valuable student data available to professors under the guise of FERPA?
  • Are universities’ internal systems contemporary or are they stuck in the 1990s? (The answer is often the latter.)
  • Do employees stubbornly insist upon using Excel as a Swiss Army knife? Or do they use powerful and interactive dataviz tools such as Tableau and Microsoft PowerBI?
  • Are universities and colleges only capturing and analyzing structured data?
  • Do employees possess the skills or even the desire to embrace analytics?
  • Are institutions’ analytics efforts university-wide or are they isolated to silos?

While exceptions abound, generally speaking I have my doubts. Sadly in higher ed, fascinating stories such as how the University of Arizona and Georgia State have used data to reduce student attrition still represent exception that proves the rule. In the jargon of Silicon Valley, very few schools seem to eat their own dog food.

Simon Says: Higher education struggles with analytics but opportunities remain.

So many questions and so few answers. Despite the growth of analytics programs (or perhaps because of it), it’s the Wild Wild West out there. I suspect that we’ll see greater progress and cohesion over the next few years.

Feedback

What say you?

Footnotes

  1. My new professor gig hasn’t changed my insomnia and moderation issues. They remain alive and well.

philanimated

Navigation 🗺️

⬅️ BACKRANDOM 🔀NEXT ➡️

Filed Under 📁



Enjoy this post? Click here to subscribe to this RSS feed or here to sign up for my bi-monthly newsletter.


Submit a Comment

Your email address will not be published. Required fields are marked *

Pin It on Pinterest