“How anything can be predicted”
by Ian Ayres
I have a nephew working on a PhD in data mining. This book has been a great help in my desperate attempts to at least have a clue about his speciality. It is written for the general reader, and is full of interesting stories about what is being done with large scale data sets. The world’s data warehouses are currently groaning with data picked up from our daily transactions. Storage is being discussed in terms you may not even have heard of... not gigabytes or terabytes but petabytes and exabytes.
Why do we need to know about this? Apart from curiosity (and those of us with data mining nephews), we are all aware of the astounding rise to prominence of the Google company. The world of massive data accumulation and “mining” is Google’s world. How is it possible that Google maps now shows constantly updated traffic density on Sydney’s roads... and why do they give this away free? One answer at least is that data is the new currency. Google gives away useful information to get more and more people using their services, thus providing more and more data for other uses.
I found this a most engaging read from the opening story about Aorly Ashenfelter who shocked wine connoisseurs by using data sets to predict the dollar value of particular vintages. The book is worth a solid end-to-end read but you can equally well dip into it randomly for a peek at this new world of digital wonder and ethical challenges.
No comments:
Post a Comment