White Paper Workflow-based Human-in-the-Loop Data Analytics

1. The challenge of data analytics The recent decade has observed a tremendous advance in data analytics. Powered by sophisticated computational techniques and powerful software tools, users interact with data analytics systems to sliceand-dice data and integrate results into their sense-making or decision-making process. Examples are abundant: scientists constructing theory from experimental observations, business operators examining financial data to safeguard and improve operations, and military officers interpreting diverse data sources for situational awareness. In all these applications, data analytics has become an essential tool that liberates users from tedious data processing tasks and allows them to focus on issues demanding more sophisticated human intelligence. Yet data analytics has its limitations. Operations are often limited to low-level processing, while high-level intelligence is scarce, hard to build, and sometimes hard to use. The challenge arises from (at least) two fundamental factors: