Data-driven Requirements Modeling: Some Initial Results with i

Requirements acquisition is widely recognized as a hard problem, requiring significant investments in time and effort. Given the availability of large volumes of data and of relatively cheap instrumentation for data acquisition, this paper explores the prospect of data-driven model extraction in the context of i* models. The paper presents techniques for extracting dependencies from message logs, and for extracting task-dependency correlations from process logs. The preliminary empirical results are encouraging.