Always Look for Bias in Data Analytics

A short, somewhat simplistic, but point-making piece in KD Nuggets on biases in analytics.  Note these kinds of biases occur in all kinds of analytics:  Descriptive, Statistical, Big Data or Deep Learning.  I have seen considerable and damaging biases in enterprise analytics decisions,  so it is good to repeat this point frequently.    Biases always exist,  you need to look for them.  They are the most common reason for analytics failure. 

Confirmation and Selection biases are mentioned as first and most important, as they often are.  I consider these together.   Also, understand that these biases exist even when you are not doing analytics,  many gut-level decisions are driven with them as well.  As the old shampoo directions used to say: Rinse and Repeat ... and make sure you know the data you are working with is clean and unbiased.