Transforming data into information

In spite of a long history of automated instruments being deployed in the water industry, only recently has the difficulty of extracting timely insights from high-grade, high-volume data sets become an important problem. Put simply, it is now relatively easy to be “data-rich”, much less easy to become “information-rich". Whether the availability of so many data arises from “technological push” or the “demand pull” of practical problem solving is not the subject of discussion. The paper focuses instead on two issues: first, an outline of a methodological framework, based largely on the algorithms of (on-line) recursive estimation and involving a sequence of transformations to which the data can be subjected; and second, presentation and discussion of the results of applying these transformations in a case study of a biological system of wastewater treatment. The principal conclusion is that the difficulty of transforming data into information may lie not so much in coping with the high sampling intensity enabled by automated monitoring networks, but in coming to terms with the complexity of the higher-order, multi-variable character of the data sets, i.e., in interpreting the interactions among many contemporaneously measured quantities.