Decades of attempts to convert this "vast amount of disparate amount of data into actionable insights" have found that it's far more complicated than you think.

Decades of attempts to convert this "vast amount of disparate amount of data into actionable insights" have found that it's far more complicated than you think. See for examples http://en.wikipedia.org/wiki/Educational_data_mining http://en.wikipedia.org/wiki/Educational_data_mining and http://en.wikipedia.org/wiki/Learning_analytics http://en.wikipedia.org/wiki/Learning_analytics for some of the well-known names for these sorts of analytics. For example, "Investing in expensive data analysis systems may not be as important as investing in ways of measuring teacher effectiveness that can identify the specific supports teachers need to improve their practice." - http://www.ascd.org/publications/educational_leadership/may10/vol67/num08/Using_Value-Added_Measures_to_Evaluate_Teachers.aspx http://www.ascd.org/publications/educational_leadership/may1... . "Value Added Measures" are one of the many attempts at applying analytics to education. The error rate and variability of that method is so high as to make it unusable. See http://dangerouslyirrelevant.org/resources/value-added-measures http://dangerouslyirrelevant.org/resources/value-added-measu... . I cannot reconcile your statement with the last 15 years of high-stakes testing, which were justified by saying they could do the analytics that you think haven't been considered. Nor with the many attempts, dating to at least PLATO in the 1960s, to develop computer assisted instruction systems, with data acquisition and analytic components. Edit: Also, "Google Apps for Education" - https://classroom.google.com/ https://classroom.google.com/