Big Data in the physical sciences: challenges and opportunities

We provide a brief overview of challenges and opportunities in the physical sciences arising from data which is extreme in volume, acquisition rate, and heterogeneity. These require novel methodological approaches ranging from data management to analysis techniques to inference and modelling, affording close links with cutting-edge research in data science. We highlight past methodological breakthroughs with high impact, showcase selected current data science challenges in the physical sciences disciplines, and broadly discuss needs of the community in the era of Big Data. We argue that the recently founded Alan-Turing Institute (ATI) is ideally positioned to facilitate, and intensify, the mutually beneficial cross-fertilisation between core data science developments and their application in the physical sciences. Concrete measures are proposed to achieve these goals, critical for ensuring impact and continued research leadership in both fields.