Reading List – 4/23
Synopsis – I wanted to start with two articles that are essentially talking about the same thing. In all bases, this is looking at how the use of “big data” or other statistical measures is leading some to blindly believe the hype of big data. The fundamental problem essentially comes down to the correlation vs. causation issues that all of you have heard me rail on for years, so I won’t go too deep. I also like to point out that like all niche groups that try to become main stream, you are seeing major dividing lines develop. On the technology and vendor side, you are seeing more and more talks about technology, platform, engineering, data “scientist” and how no company can compete without “big data”. On the other side is the push back pointing out that none of this is new, interesting, and that in almost all cases people are claiming things from big data that cannot exist. I would strongly suggest that if you get the time, that you check out a few of the articles mentioned in the big data roundup Forbes article. If you want a better way of viewing the world of big data (even if it does not go deep enough in establishing ground rules) I would focus more in the NYTimes article:
Synopsis – Dan Ariely is one of the best current thinkers on human nature. In this RSA video, it breaks down the rationalization and cognitive approach to dishonesty. The key reason I am sharing this, besides just a look at human behavior, is the same logic works to explain why people who have been doing the same thing over a long period of time believe they know everything about what they are doing. I also think the look at the scale and impact of small rationalizations and how massive an impact it can have to an economy and for a business:
Synopsis – There has been some recent talk about whether it is better to use P-Score confidence measures or Bayesian approaches. This looks at the many problems of “statistical confidence” and how it leads people astray. I know this is hardly a new topic, but I do think the look at different ways of thinking about the same problem can help everyone really focus on what they are trying to learn when they use similar types of statistical tools. I also think the example problems help make the issues clear:
On a related note to all of the above, I wanted to make sure that everyone saw the news from the last few weeks on how excel errors were missed and lead to massive economic theory changes in the last few years:
Synopsis – For our worst article of the email, I wanted to send this amazing breakdown of analytics for business leaders. I am less interested in the bad content associated with this one and I am far more interested in how this is an obvious attempt of someone who is trying to stay relevant in an industry and trying to show their “expertise” without any real knowledge or functional skill on the topic. This is vital to read and understand as the key problems we face are changing this type of mindset to go in completely different directions and to think differently:
AA
TL;DR 