I just finished this book, Leading with AI and Analytics: Build Your Data Science IQ to Drive Business Value. Northwestern University Kelloggians (I don’t care if that’s not the proper demonym, it is now) Eric Anderson & Florian Zettelmeyer offer a good introductory text for anyone interested in using artificial intelligence and analytics in their business enterprise, particularly a nontechnical manager. They call this “AIA,” which always throws off the part of my brain that has background in architecture and construction (so I guess I’ll call it AI&A to avoid confusion with the American Institute of Architects). And, they say, it can positively transform your business enterprise– so long as you understand some key principles, ideas, pitfalls, and a bit of managerial theory behind it. I read this to follow up on a frustratingly sparse curriculum in a business intelligence course in my MBA program, thinking it might give me some more technical background that was lacking from the course.
It did, but only barely. First, the good stuff: Anderson & Zettelmeyer provide a really valuable conceptual framework for how to improve what they call your Data Science IQ, or DSIQ. This is a gimmicky term here– business people love gimmicky terms- but it’s important to illustrate how any BI or analytics framework is only as smart as the person who is designing it, controlling the inputs, and interpreting it. The book touches on vital issues such as the distinction between data integrity and the notion of “truth in data”– the idea that there can be multiple, competing truths in most business scenarios, and they can only best be understood through multiple angles, parameters, and, most importantly, a manager who understands the inputs and outputs.
Different Types, Pitfalls, and Questions of Analytics
The authors provide a good background in understanding the distinctions between different types of analytics (exploratory, predictive, or causal), and explain how to work through some of the managerial pitfalls that hamper the success of AI&A implementation (data siloing, poor quality integration of disparate and legacy business information systems, inability of data people to talk about what they do and an inability of managers to understand). They also give any manager a good framework in terms of eight key questions the manager must ask when approaching an AI&A project, making sure the reader knows about the importance of asking the right questions and figuring out how to get the job done. (Do you need every one of these eight questions? I dunno, I still have a course left in my MBA, so I’m not an expert. Maybe I will unlock the secrets in this last course and be able to answer in the affirmative).
The Less Illustrious
The more frustrating components: I was really hoping for a more technical dive into things like data sourcing and processing as far as “big data” stuff like data cubes or OLAP/OLTP– the idea of data having dimensionality beyond just a simple spreadsheet. I was also looking for much technical background more on neural networks, random forests, gradient boosting, decision trees, and analytics languages. Languages– like R or SQL- increasingly seem a given for anyone doing anything vaguely analytical or quantitative in this modern economy. Statistics is definitely a given if you’re interested in AI&A, so I don’t mention that in this list, but they also don’t get too far into this, either– beyond some general ideas about things like data distributions and the use of regression analysis in predictive analytics.
AI & Analytics, or just statistics?
Indeed, stats underlies a lot of this, because a lot of machine learning involves getting computers to perform repetitive analysis to figure out certain outcomes based on certain inputs. So, wouldn’t hurt none to get a book on stats, too! I am not very good at math or science, but I do appreciate some background on the stuff. Sort of like how I don’t need a pharmacy degree, but I do appreciate reading the Wikipedia page on a drug to understand what it does to calcium channels or GABA receptors in my brain or what have you (leave the rigorously technical processes to the scientists themselves, but make sure you know how this stuff works).
…or actually just, like, strategy?
Like many a business text whose authors are better at producing content than they are at actually writing that content from a stylistic or narrative standpoint, the book is a bit redundant at times. Part of this might be valuable for the lay reader, dilettante, or novice, but a lot of it will feel repetitive if you’ve already read a few business texts, especially as far as strategy. Business strategy is principally concerned with asking the right questions, and this is a fundamental part of BI and analytics as well.
Most of the insights the authors are offering are therefore not really AI & analytics insights as much as they are insights about business strategy in general– because they focus on asking the right questions about information, objectives, and how best to use the former to achieve the latter (and how to avoid things like false or otherwise befuddled causalities, or things like confirmation bias– again, just as much about stats as about strategy, and none of these are very technical concepts).
Anyway. It’s a good read, but I kind of wanted something to blow my mind, and this ain’t it. YouTube these days is full of crash courses on AI&A and machine learning specifically. This book will touch on these things and give you some vague background, but that’s about it. Maybe they want you to go on YouTube before or after?
But then like, why did I spend all the money on this darn book? It’s sort of the MBA school of thought, which is “you don’t need to understand how any of this stuff works, you just need to know how to manage people who do!” Which is like, cool, but that wasn’t worth paying for my MBA (or the book, necessarily). I might have a different perspective if I had no background whatsoever, so just bear that in mind if you buy this. It’s a pleasant read, overall, and not bad, just not really about to rock anyone’s world.
Check out Leading with AI and Analytics: Build Your Data Science IQ to Drive Business Value. Or, like, YouTube.