Engineering Analytics for big data

Efficient Analytics are at the heart of any nontrivial next generation computer application. But how can we obtain innovative Analytic solutions for demanding application problems with exploding input sizes using complex modern hardware and advanced Analytic techniques? This tutorial proposes Analytics engineering as a methodology for taking all these issues into account. Analytics engineering tightly integrates modeling, Analytics design, analysis, implementation and experimental evaluation into a cycle resembling the scientific method used in the natural sciences. Reusable, robust, flexible, and efficient implementations are put into Analytics libraries. Benchmark instances provide further coupling to applications. We begin with examples representing fundamental Analytics and its structures with a particular emphasis on large data sets. We will also give examples of future challenges centered on particular big data applications.