I lead Trexin’s Analytics Capability, but as usual it was not a linear path to get here. Decades ago, I wrote code and a term paper as part of an independent study course in Artificial Intelligence that I developed with my college advisor. Not surprisingly, I failed to come up with a conversational agent that fooled anyone or even worked, but I learned a lot and enjoyed trying. Progress in AI was slow, and I abandoned the idea of pursuing it.
I now have a degree in Computer Science and decades of experience developing software and participating in every aspect of the full software development lifecycle. I sincerely loved a lot of that work and working in a team to create great new products and systems, but at the same time, rarely felt a true passion for it as I had for my study of AI.
About ten years ago, I was writing Java code at work when a friend who was working on doing some spreadsheet-based analysis walked up to my desk. As we talked, I realized she was going to have to spend days doing her task manually, while I knew I could write some code in an hour or two that could do the same thing – so I did.
When I had finished, we started talking again about the data she had compiled, and I realized that I could write more code to do analysis that would be difficult or impossible to do in Excel. I started writing more Java code, but another friend pointed me to R (the open source statistical language). I used R to create some relatively sophisticated analysis in a day or two, and never looked back.
In the last ten years, I’ve worked in the trenches of analytics and learned a lot, mostly about how analytics initiatives and programs fail in corporations, but also how it can literally save lives when done right. It’s hard to think of a better reason to do something than that.
This blog is called “Pragmatic Analytics” both as a small homage to The Pragmatic Programmer (a great book by Andrew Hunt and David Thomas) and to reflect my focus on making pragmatic choices about how companies should and shouldn’t do analytics, grounded in years of experience. The book blurb starts “Straight from the programming trenches, The Pragmatic Programmer cuts through the increasing specialization and technicalities of modern software development to examine the core process…” To me, this mindset captures a lot of what I want to say about analytics.
An idea I had many years ago crystallizes this for me. I was toying with the idea of creating a technology startup and calling it “Mid Tech”. Companies were spending a lot of time implementing the latest hot technology without understanding why, and consequently were losing money. In their pursuit of high tech, they were ignoring very valuable technologies that had been in vogue in the past but had been overlooked when the next thing came along. I wanted to call my firm Mid Tech to reflect an approach more, well, pragmatic than the search for shiny things. Not a great marketing idea, I’ll admit, but it does capture the idea that one has to be grounded in reality when considering how best to implement new ideas.
A cartoon meme recently made the Twitter rounds showing, in its four panels, a person looking at a crack in a wall labeled “statistics”; in the next panel, he’s putting a frame around the crack. The third panel shows the now-framed crack labeled “machine learning”, and in the final panel, the label now reads “Artificial Intelligence”, and there’s a crowd of people in front of it.
There’s a certain amount of breathlessness around analytics now (also including technologies I’m passionate about, like Machine Learning and Artificial Intelligence), and while I’m as starry-eyed as the next person about the incredible progress and possibilities, I’m also pragmatic in considering all of the less-glamorous realities of bringing value to companies using data science. This is the intersection I want to explore: how to implement analytics (including those most exciting ideas) in a pragmatic way to maximize the value companies get from their investments of time and money.
My next post is called “Why Another Analytics Blog?”. I’m pretty sure you can guess what it’s about. In the meantime, thanks for reading. I’d love to hear your comments about what possible reason there could be to launch another blog about analytics!