The evolution of cloud architectures, and of industrial data-driven decision making. We will cover subjects ranging from hypothesis-driven queries to opportunistic data analytics, while dealing with issues such as data inhomogeneity, poor labeling and the curse of dimensionality.