Data analysis

In speaking of data analysis and pipeline development, we refer to the branch of cosmology that bridges the gap between the use of data as a tool to inform theoretical models and efforts devoted to hardware setup and raw data taking. Cosmology deals with extracting information on the large-scale Universe from observations. It is a peculiar discipline, since the system under study and the laboratory where the investigation takes place, are the same. This makes cosmology extremely stimulating, because the techniques it employs are continously updated, based on the most recent developments in the field. A clear example is given by measurements of the cosmic microwave background (CMB) radiation: starting from the first experimental evidence of its existence, more than 50 years ago, to the current measurements and, further, to the future observational campaigns, experimental apparatuses, as well as the algorithms for data taking and analysis, have evolved in a rapid and incessant manner. CMB observations have reached an extremely high level of sofistication, leading to precise measurements that have allowed to assess with equal precision the validity of theoretical models. A similar pattern has been followed by observations of the cosmological large-scale structures (LSS), performed through increasingly sophisticated surveys, that have been complemented by continously developing analysis pipelines.

In both cases, it is of the paramount importance to get to know the finest details of the experimental apparatus, and to be able to predict (and possibly mitigate) the potential impact of systematic effects on the observations and on their interpretation. From this stems the strategic importance of being able to simulate, as a whole, the instrumental characteristics of the experiment, the observational campaing, and the following data-analysis step (end-to-end simulations) even before the experiment sees its “first light”. This simulation activity is, to date, an essential element in the design and development of a cosmological mission. In this context, computing science is key to the research in cosmology.

The cosmology group in Ferrara has developed a decades-long experience in this field. Members of the group have contributed to the success of several CMB missions, the Planck mission among all of them, playing important roles within the corresponding experimental collaborations. Currently, the group is involved in several scientific collaborations, including the future CMB missions LiteBIRD, LSPE, Simons Observatory, CMB-S4, and the future LSS survey Euclid. The group activities are focused on the following topics:
– development of end-to-end simulations, with emphasis on the modeling of instrumental systematics, in order to propagate the uncertainties related, for example, to the characterization of detectors or polarization modulators (half-wave plates);
– development and validation of data analysis pipelines, including building algorithms for the statistical analysis of the data and for their reduction to final analysis products (for example, maps to power spectra).

These activities are closely related to, and represent a natural complement of, the more theoretical research lines carried on by the group. The educational and professional paths of the members of the group all have in common, within individual differences in inclination, a strong degree of multidisciplinarity. For this reason the cosmology group gives great importance to the ability to move between theoretical, observational, and phenomenological activties.