ABSTRACT
Context: Process performance in Information Technology (IT) Service Management is traditionally measured through process performance indicators (PPIs). These indicators allow a quantitative evaluation of the efficiency and effectiveness of a business process. However, despite the importance of PPIs, it is unusual to find clearly defined and complete sets of indicators in actual organizational environments. The lack of access to PPI definitions sometimes makes it difficult to evolve new proposals or improve existing techniques for the definition and analysis of PPIs.
Objective: In this chapter, we present the PPI Dataset, a set of PPIs defined in natural language, related to the IT service management processes of several public organizations in Spain. This dataset has been used as a basis for the development of techniques for modeling PPIs using different notations as well as for the transformation of natural language PPI definitions into others amenable to automated computation, applying natural language processing and matching learning techniques.
Method: The PPI information was obtained from the Service Level Agreements (SLA) of the technical specifications of the request for quote (RFQ) that these organizations have made. A preprocessing of the data was carried out to verify the completeness of PPI definitions. The data was originally provided in Spanish and was translated for dissemination purposes.
Results: We obtained a total of 102 indicators associated to different processes. This dataset has been used in several research articles, two of them focused on the automatic transformation of PPIs defined in natural language to others that are defined in a structured way to facilitate their automatic calculation, and one oriented to the definition of PPIs through graphic notations.
