This is my first book in my #52in52 attempt (to try and read a book a week for a whole year). This is going to be reading for pleasure, so whilst this title was non-fiction, future books will be a mix of fiction and non-fiction and not necessarily professional development focused (though I would argue increased reading for one’s own self will have non-direct professional benefits)
I should say outright that I really enjoyed ‘Everybody Lies’ by Seth Stevens-Davidowitz and would recommend it to anyone interested in this area – and who wouldn’t be!
I have some very basic experience I this field – having in the past used market segmentation tools (ACORN from CACI) to focus offers but this didnt dampen my enjoyment and I don’t think you need any prior experience/knowledge as it’s written in the informal academic style that is informative but not exclusionary.
The basic premise is one of introducing and espousing the concept of big data without the need for a background in data science and that of a ‘digital truth’.
The author uses (intentionally?) attention grabbing topics and examples especially earlier in the book such as racial prejudice and voting patterns and po*nhub preferences (I’m not one for censorship but I suspect there would be some spamming involved)
It may be my public health background but I was most interested in the randomised trials and A/B testing section. It was fascinating to see how big data and especially big data organisations have adapted certain scientific principles to improve their offer or services.
The book gives examples of how big data is most effective where strong systems are not in place and gives a fantastic example in linking health searching to poorer cancer outcomes (pancreatic I think…) and how big data can be used to highlight a particular population. Again, with my Public Health background, it would be interesting to see a full cycle of big data problem solving – not just the issue/population highlighting but an intervention at said group and evaluation of it’s effectiveness. I am sure this has already been done/in process though.
As the book drew to a close, it did look at the potential pitfalls of big data usage on both commercial and public sectors. Issues such as ‘correlation not equalling causation, generalisation and the curse of dimensionality. Don’t trust your lucky coin…
It’s always nice to reach same conclusions as the author (and the sign of a good read) so the discussion about how big data is underused and wonderful, yet not *the answer* rather an undervalued part of the wider answer wasn’t unexpected and the point about data science moving from social to ‘proper’ science was strongly made.
My final thought would be I hope the author enjoyed his beer (if you read to the end you’ll know what I mean)
As you can tell, I really enjoyed this! Now onto my next book…