During TEDx at the Rotterdam School of Management Scott Mongeau, manager analytics for Deloitte, gave an interesting talk about how we can be tricked by the analysis done with big data. He explains that models used in big data analytics should be driven bye the data and not the other way around. Essentially all models are wrong, but some are useful and data analytics can help solve some of the problems that we have currently, but it is important to have better diagnostics and semantics.
Thill will ensure that the models used are sound and can better describe computers what we are trying to achieve. He says that we should not become so focused on the engineering part of big data analytics that we forget the social context, which is important to understand as well in order to create true insights.