Summer time is a good time for reading…
Introduction to Bayesian Statistics by William M. Bolstad
A clearly written, enjoyable read that starts from first principles. (Don’t you hate it when a book assumes that you already know why you should be reading it?) Bayesian statistics are extremely hot right now, so it’s been a priority of mine to shore up my lack of knowledge in this area. In my undergrad and grad years during the 90s, many computer science departments placed very little emphasis was on basic stats. Mine was one of them, and I missed out. It’s a shame, because these days stats are at the cornerstone of analytics applications driving many emerging businesses (as well as supernovae such as Amazon and Facebook). Bolstad’s book provides the justification for Bayesian stats (versus the stats you learned or snored through in school) as well as the theoretical and practical depth to turn blog buzzwords into concrete concepts. It’s lead to many enjoyable weekend afternoons screwing around in R. The pricetag is hefty, but if you can get a copy it’s worth it.
Competing on Analytics: The New Science of Winning by Thomas H. Davenport and Jeanne G. Harris
This is an outstanding survey of how today’s businesses use analytics to drive results. This book is not technical – it is targeted at business decision makers who are trying to figure out why they’re getting their asses kicked by the guys down the street. The authors are not mathematicians (so far as I can tell) but do a great job of translating the often arcane lingo of the operations research and stats world into the often arcane lingo of the MBA graduate. Kidding aside – speaking to both worlds is a very useful enterprise. For these reasons this book is suitable for those interested in analytics because of its business value as well as for practitioners.
The authors do not explain how to implement analytics-based tools and processes. That is not the point, and in any case it would be silly to think that a single book could provide all the answers. Davenport and Harris outline the full lifecycles associated with adopting analytics and applying analytics. Adoption occurs in stages – from an initial awareness, to aspiration, to competency (perhaps at a departmental level) to becoming a fully capable analytic competitor. The authors claim that enterprise-level adoption requires top-level management support – a claim that is supported by my own experiences (sadly not as a top-level manager). Analytics is sometimes confused with (or reduced to) enterprise data management, spreadsheets, or SQL queries. Analytics is something much broader, and thankfully Davenport and Harris describe the extremely important role that operations research plays in analytics. Operations research is in some sense the highest form of analytics in that it is aimed at producing decisions. (As described in a recent article in Analytics.)
Competing on Analytics is a quick read but will keep you thinking. The Kindle version is quite inexpensive.