Tech - Book review 'Designing Great Data Products'

Posted by Rajan Manickavasagam
May 15th, 2012



Rating: 3 out of 5

Image: O'Rielly

Many companies and commentators highlight the growing data that is created and consumed everyday. I was naturally curious to understand more about how companies can use all this data to their advantage. I came across this book - 'Designing Great Data Products' in O'Reilly. The title sounded great, but to my disappointment, it is a short book.

I anyway went ahead and read the book.The authors of the book are folks from kaggle, a company specialising in statistics and analytics. The book mainly talks about an approach called 'DriveTrain'. This  approach is about using data to drive actions.

The 'DriveTrain' approach consists of 4 steps -

  1. Define a clear objective - I like this one. Please refer my previous post about problem definition.
  2. Levers - Identify the systems, processes etc that can be controlled.
  3. Data - Is there existing data that can be used? Or does additional data need to be collected?
  4. Models - In my opinion, all of the above steps are critical to choosing the model. The chosen model would then be used to solve the problem.

The book then takes us through an example of how the above approach was used in the insurance industry. The approach was used to allow insurance companies to price their policies better. The authors also explain how this approach could be used in other scenarios, such as recommendation engines, optimizing lifetime customer value, traffic management and so on.

I would recommend this book as a quick introduction into the world of data analytics. The 'DriveTrain' approach is pretty generic enough to be used in various scenarios.

Disclaimer: I am a member of the O'Reilly Blogger Review program.

Comments

Popular posts from this blog

Cloudera Quick Start VM in Hyper-V

Tech - Check your Salt - Part 3

Prime Candidate for Analysis