RADAR, A FRAMEWORK FOR AUTOMATED REPORTING Antonia Azzini 1 , Nicola Cortesi 1 , Amir Topalovic 1 and Giuseppe Psaila 2 1 Consortium for the Technology Transfer – C2T, Via Marengo, 22 – Carate Brianza (MB) Italy 2 University of Bergamo, viale Marconi, 5 – Dalmine (BG) Italy ABSTRACT Large companies and organizations periodically feed their information systems with large data flows. Apart from the classical operational activities, they are called to prepare aggregated reports to send to institutions and rating agencies. Unfortunately, organizations typically suffer for the lack of integrated data and for the lack of a standard data dictionary. The presented approach aims to tackle such a probl em by building a “bridge” between employees that need to specify how to generate reports (on the basis of concepts and terms typical of the application domain) and the information system that stores the data to query and aggregate in order to automatically produce reports. The implemented framework, RADAR (Rich Advanced Design Approach for Reporting), moves from the notion of Operational Data Store, and it is posed in the middle between an ontology (of concepts and terms) and the actual operational (and relational) schema of source data. Then, in the defined schema allows for giving a high-level view of such source data, based on concepts described in the ontology for a specific application domain. KEYWORDS Knowledge Representation, Automation of Services, Data Integration, IT Services 1. INTRODUCTION Current information system technology is able to store and process a huge amount of data. In case of large corporations or institutions (such as financial institutions, multinational manufacturers, central banks, etc.) information systems are fed with large data flows that are provided periodically (e.g., daily, weekly, and monthly). Database administrators and system integrators have been developing solutions to efficiently receive and manage these data flows, by optimizing loading processes as much as possible. Apart from traditional operational activities, collected data are essential for reporting activities. In fact, regulatory bodies (such as central banks) and national offices of statistics as well as rating agencies ask companies and institutions to periodically deliver reports illustrating the current state of their business. The reasons are manifold: they can vary from the need to monitor trends of economy, to the need to provide markets with transparent information concerning the current economical state of companies. Nevertheless, companies might need to prepare reports for internal use, for example for auditing and for decision-making. Typically, organizations suffer for the lack of integrated data and for the lack of a standard data dictionary; then, significant manual intervention to write queries and create reports is necessary (Browne, et al., 2019). Moreover, employees who are responsible for producing reports in general are not aware of technicalities concerning how data are stored and how they can be queried; consequently, they have to interact with technicians that also have to be experts in the application domain, so as to understand requests made by employees (these professional figures are very rare). Clearly, there is a gap between users and systems that can be resumed by the following questions: “What is the meaning of data? How could they be interpreted to meet the requirements? How could people, who are not database programmers, define reports?”. To further complicate the scenario, terms and concepts are often informally defined by international organizations; furthermore, the meaning can vary in specific communities. Specialists use a lot of complex interconnected concepts, including those defined by international standards and systems, developed to provide, e.g., global financial services. Clearly, a common comprehension of terms and their semantics is essential to prepare reports: knowledge about terms and data is crucial to define reports effectively and efficiently. ISBN: 978-989-8533-95-1 © 2019 54