Visual analytics algorithms for multidimensional astronomical data

The growth rate at which radio astronomy data is captured is accelerating thanks to recent developments in phased-array feed technology that increases the field of view of radio telescopes. With this improved instantaneous field of view, the number of known neutral hydrogen (HI) sources is set to increase rapidly with SKA precursor’s wide-field surveys. This large field of view is inherently linked to large data volume and data rate. To be able to rapidly process large datasets, and explore information hidden within this data, new algorithmic approaches and analytical tools are required. Graphics shaders—algorithms interpreted at runtime to process and visualise data in real-time—can alleviate bottlenecks caused by today’s data challenges and hardware limitations. On highly parallel hardware like graphics processing units (GPU), thousands of shaders can be computed in realtime. The graphics pipeline of modern GPUs allows for the development of custom vertex, geometry and fragment shaders that can be exploited to improve on astronomy-centered analytical visualisation practices. In this talk, I will present 3D statistical visualisation algorithms that can be used as building blocks in a streaming framework to rapidly explore data ranging from hyperspectral and multiwavelength imaging studies of galaxies, planetary nebulae, and time-domain astronomy in an intuitive manner.