The Data Analysis Working Group considers developments and approaches that will make the analysis of HEP data towards final physics results faster and more scalable in the future. One new approach to be considered is the development of expressive functional interfaces for analysis. The meshing of HEP specific tools with those found in the larger data science community is another area of interest, where database approaches or formats optimised for high-speed processing may be profitable. The development of test beds for high performance analysis clusters, that may speed-up turn-around and efficiency is a major area of R&D. The group should help coordinate activities and communication here between experiments and IT experts. The integration of such systems into the end-to-end data flow, with caching and distribution policies, are important. Optimising the actual content of analysis formats is a boundary condition to discuss with physics groups, but can bring major gains.
Convenors: Paul Laycock (Belle II), Danilo Piparo (ROOT), Andrea Rizzi (CMS)