I’ve started to try to actively learn from books rather than just read them. I’m posting my book summaries online to keep myself accountable to my “book a week” goal and hopefully to encourage others to read more too.
This Week’s Book Title [Week 34] : Weapons of Math Destruction – How big data increases inequality and threatens democracy
by Cathy O’Neil
Paperback. 218 pages of content + 34(!) pages of references/citations (I love that!)
The book looks meaty…there is a lot of text per page and not much space in the margins. The paper is nice and white and the font and text size are easy to read. Despite it being a book about maths, I don’t see a single formula when I flick through the book (whew…reading formulae isn’t my strong suit).
The book is very easy to read and lays out in plain language how data-driven models effect a lot of areas of our lives – college applications, credit scores, insurance premiums, etc. , and makes complicated financial systems understandable without having to know maths.
I’m finding myself looking forward to reading time and not missing watching so much TV. The pessimist in me wonders if that will last once all the TV shows come back in a few weeks time. Anyway, back to the book…
The author introduces us to the concept of a Weapon of Math Destruction by talking about what a model is (an abstract representation of some process, be it a baseball game, a supply chain, etc) and by giving examples of how those models can be wielded as weapons – often against the most vulnerable members of society.
The book starts out with an example of a good model – baseball – where the model is open and clear for anyone to examine. There are constant inputs to the model during the baseball season. The model is tested, adjusted where it failed and tested again. It is a living thing.
In contrast, many other models described in the book are created based on a set of chosen inputs (that may be biased or completely miss other important inputs) and then never adjusted with new data i.e. they grow stale. How the model was built and how it operates is usually not disclosed to the people affected by the model i.e. it is opaque. The success of a model is usually based on profit or efficiencies introduced, not it’s benefit to society.
Throughout the examples of models, there are mentions of regulations governing certain things e.g. credit scores and how companies can bypass regulations by creating escores with data available online. The author suggests that it is time for the law to catch up and protect people against such models.
As well as legal protections, the author suggests that modelers should abide by a code of ethics and cites an oath drawn up in “The Financial Modeler’s Manifesto” by Emanuel Derman and Paul Wilmott in 2009. It reads:
- I will remember that I did not make the world, and it doesn’t satisfy my equations.
- Though I will use models to boldly estimate value, I will not be overly impressed by mathematics.
- I will never sacrifice reality for elegance without explaining why I have done so.
- Nor will I give the people who use my model false comfort about it’s accuracy. Instead, I will make explicit it’s assumptions and oversights.
- I understand that my work may have enormous effects on society and the economy, many of them beyond my comprehension.
The few paragraphs above do not convey how important I think it is that if you’re studying or working in the area of Data Science, then you need to read this book.
I knew that models drive a lot of things in our society – from what news we see to what price we pay for groceries – but to see it all laid out in examples like this gives an awesome insight into the effects – good or bad – that models have on us.
This book has had what I think is the desired effect on me – when I create models in future, I will make sure to think of the wider consequences of doing so, rather than just getting the error rate down another percentage point.
Next week’s book
In case you want to read along with me, here’s what I’m planning to read during the upcoming week: 59 Seconds: Think a Little, Change a Lot by Richard Wiseman ( not sure whether to categorise it as Personal Development or not with this quote by Derren Brown on the cover: “Excellent! A triumph of scientifically proven advice over the myths of self-help. Uplifting and long over-due”)
For anyone who’s wondering, yes…I’m STILL waiting on Ready Player One (Science Fiction) and The Miracle Morning (Personal Development) from the Library, Last time I reserved a book it came through quite quickly. I suppose these are really popular at the moment so have to go through a few other people first.