Logo: to the web site of Uppsala University

The capability to efficiently handling and analysing data streams in industrial processes and industrial cyber-physical systems (ICPS) is critical for digitalisation and renewal of current manufacturing industry. A key problem within this context is to provide scalable capability to collect, process, analyse, and visualise data streams to support these ICPSs. The status of these systems continuously changes, and analysts must understand and act upon such changes often in real time. Visualisation tools are increasingly used by analysts to get insights from these changes, but inconveniently nowadays, analysts have to store the data from the Industrial Data Stream to a data storage system and then use some visualisation tools to be able to visualise the data to help them understand and act promptly. In this paper, we propose to integrate visualisation and analysis primitives transparently into the query language of the Data Stream Management System (DSMS) and we show a proof of concept by a successful integration of an operator that can execute query-based visualisation methods that support processing of continuous numerical queries over streaming data in industrial analytics applications. We will also show how we are benefiting from the meta data in DSMSs to perform dimensionality reduction in real time to identify a two, three, four, or five-dimensional representation of the numerical query results on the data stream which preserve the salient relationships in the results and how the operator can suggest the most appropriate visualisation of the data to the analyst.