ABSTRACT
This chapter presents a new perspective on Big Data, surveillance and crisis management by adapting approaches from the fields of financial sociology and from the emerging field of the social studies of valuation. It deals with the social construction of the algorithms and the implied establishment of specific orders of value which are mandatory steps in creating emergency management systems. Event detection and tracking are key elements in emergency management, as they not only allow the finding of new and emerging crises, but also monitor their evolution. The chapter describes a generalized and partly simplified overview helps to understand the specific characteristics of different algorithmic practices. The algorithms search for subjacent patterns of user behavior and word similarities, which indicate with high accuracy, precision and recall the corresponding results in the training data. To refine the algorithms and developers and researchers modify parameters, "they play around" to bring the algorithms closer to what has been classified as relevant by humans.
