If you have a theory, models and methods you are lucky. If yo know a little about dependencies and distributions, you apply statistics. If you have nothing, you are forced to use data driven methods. If you are intelligent, you combine them. If you like to optimize resources, you take a platform that supports exactly this.

SW Developers Should Learn Machine Learning
The Exabyte Revolution
Economics Needs Data Mining
mlf Goes OS X
Econometrics Learning
The AI Revolution Is On
The Blank Swan Of Metal Treatment
The Next Scientific Revolution
Forecasting or Predictive Modeling
Being Tracked
Industrial revolution of Data
Huginn and Muninn
The 7 Deadly Sins
Beware Wind
I Saw It With My Eyes
One Size Does Not Fit All
Mlf 2 Grid Enabled
The Petabyte Age

Use Cases

How Much Water
Predicting The Presence
Understand Your Processes Better