Big Data, Big Dupe: A little book about a big bunch of nonsense

I found this book to provide very interesting insights regarding how big data can be used in a useful way in today's world. Big data is not a panacea but rather a good tool, in the same way as data has always been.

In this book Stephen Few seems to be fighting a fight that nobody perceives as necessary or framed in the manner that he frames it into. The book can be summarized in 2 points: 1) Big Data is just data. - He argues the above because in his view there is no accepted point in time or measure from which experts have started agreeing that from that point in time or measure, data became big data. He argues that the push of Big Data comes from tech vendors who want to sell their softwares and storage space. *In my view, his point while it could be philosofically mostly valid than not, it is in practice very much irrelevant and could even be falsified by establishing an arbitrary measure from which data can be called Big Data. Indeed, running some statistics and a few graphs 60 years ago with the available data and tech IS much different in scope and size compared to the practice of data analysis that we have today. Stephen Few discounts completely the fact that technology helps and the volume of data available makes a difference. To use one of his arguments, if many years ago you could extract, from example, 10% of insights out of the available data in 1 month, today even if we could keep the 10% insights' extraction, this would be a percentage of a larger volume of data and the time for achieving the result would be much shorter. 2) Correlation is not Causation. He argues -and cites some books on the matter-, that through what he perceives as the fraud of Big Data, certain individuals are pushing for the abandonment of Causation in the pursuit of mindless Correlation between variables in large datasets. *I have graduated in Economics and Finance, and I am currently in post graduate education studying Business Analytics. In every quantitative course that I have ever taken, NOBODY HAS EVER TOLD ME TO TRADE CAUSATION FOR CORRELATION. I also don't feel that this is the trend anywhere else, I have only found this warning in Few's book. Nevertheless, while I am grounded in the pursue of causation, a correlation matrix within the variables of datasets who have lots of records, is just another tool of analytics and it MAY indeed, sometimes, provide useful insights. In conclusion, I am rating this book poorly because I don't perceive the threats that Few's talks about in his book as legit. What I mean is, while I mostly agree in principle with what he says, I don't see like at all, that people started doing what he is warning against. In this book Few reminded me Don Quixote fighting against windmills.

Come contraltare a <i>Caos quotidiano</i> che aveva tessuto le lodi della non-strtutturazione ho preso questo libretto che ha una tesi completamente diversa: i Big Data non sono altro che l'abbindolamento che ci fa chi vende hardware e servizi di rete. Per amor di completezza, Few con i dati ci lavora; la sua tesi però - esposta in capitoli dai titoli esplicativi "Big Data, Big Whoop", "Big Data, Big Confusion", "Big Data, Big Illusion", "Big Data, Big Ruse", "Big Data, Big Distraction", "Big Data, Big Regression", è che in realtà non c'è nulla di davvero nuovo, nemmeno la grandezza relativa dei dati in questione; quello di cui abbiamo bisogno è avere persone in grado di comprendere i dati, e non credere che le macchine possano fare tutto da sole. Quello che funziona in realtà non sono i Big Data, ma per esempio il machine learning. Generalmente io sono d'accordo con Fry, anche se non arrivo alle sue posizioni talebane di un movimento Slow Data. D'altra parte, il penultimo capitolo "Big Data, Big Brother" dimostra che questi dati vengono usati eccome...

In his latest book, Stephen Few states what few are willing to say out loud - "Big Data" is nothing more than a marketing sham. From the initial chapter highlighting how no one can even adequately define "Big Data", the tone of this book is clear; the emperor has no clothes and we need to wake up. For any of you who work with data sets on a regular basis, this book is a must read. Save your companies millions of dollars by not chasing a fad. Focus on what's real - DATA. Not toys, not tools, not hype. Nothing has changed, data is data is data. I truly enjoyed his highlighting of the contradictory messages other authors and 'data experts' have published. As a data professional myself with over two decades of experience, this book is a treat and a validation. No matter what new buzzwords marketing comes up with, Few reminds us that it's people, human insight, and focus on the right data (regardless of the volume or velocity) that counts. Follow the advice given in this book and free yourself from the hype. For the tl;dr crowd: Read this book and save yourself a ton of headaches!!

This is a wonderful book that in a sense debunks the whole "big data" catch word. If you are a manager type this is a must read!