Bridging high performance computing and big data analytics with end users
By bridging high performance computing and big data analytics with end users, the Horizon2020 funded CYBELE project aims to produce technology that can revolutionise farming, increase food supply, significantly boost animal welfare and bring social, economic and environmental benefits. A key part of the project is to ensure that these state-of-the-art technologies are accesible and useful to the users, who might not necessarily know about computing at all but wish to make decisions based on data. Data visualization, or the graphic representation of information, is one of the main tools at our disposal. In a previous post (https://www.cybele-project.eu/post/designing-adaptive-visualisation-tools-for-cybele), we talked about how the visualisation team at Barcelona Supercomputing Center participates in CYBELE with the creation of an application that will let data scientists and analysts understand their results and create reports that can be used by farmers to make decisions. But how do we make a visual platform that can ingest large quantities of data and is flexible enough to accomodate the many kinds of uses that CYBELE aims to fulfill? The visual application that we are building connects automatically to the CYBELE components that produce and analyze data, so that the user never has to worry about the size of the data, or even of updating files. With that connection, users can configure different types of data visualization graphs (for example, line charts, pie charts, maps, scatter plots). Each chart can be configured with multiple options for text, annotations, colours, and so on. On the final step, the user of CYBELE can prepare dashboards or reports to send to the final users (e.g. farmers). These are collections of charts that need to be read together in order to fully understand a problem or conclusion in all its dimensions. The CYBELE visualization platform relies on extensive user experience research and testing, not only for its interface components, but also for the data visualizations that are created as well as the reports. In this case, we have incorporated all current research on how information can be layed out visually to convey different types of messages. The outcome of CYBELE will be an integrated platform that harnesses HPC and data analytics infrastructures for people that are not software engineers or ICT’ers but data analysts or advisors. The visualization creation tool is the component that will let them create impactful solutions for the agriculture and livestock sector.