The points in the article are important to consider (sampling bias etc etc) but they apply to ALL attempts to analyze data for patterns. For me the utility of

The points in the article are important to consider (sampling bias etc etc) but they apply to ALL attempts to analyze data for patterns. For me the utility of big data is simple - more samples. Even if the effect is small you can find it and have more faith (word used on purposed :}) it is not noise that has escaped your significance test. The dataset may be biased. That part requires wisdom to know and honesty to say what the limitations of the analysis are, but it is by no means restricted to "big data". But the article as a rant against buzzwords is something I can get behind.