3.2. Fundamentals

In this section you will learn the basic concepts underlying an analysis at Belle II, starting from how the data acquisition works and ending to the description of the most common analysis concepts.

The workflow that goes from the data taking to the publication of any measurement in High-Energy Physics (HEP) experiment is quite complex. The data analysis itself involves multiple steps that can take months or even years, and this comes on top of the time required to take the data, process them, and produce the corresponding simulated events. While the details of all these procedures can be extremely complex and tedious, the overall picture is simple enough to be fitted in a human-readable scheme:

../_images/grand_scheme_of_hep.png

Starting from the end, you can see that the input to the analysis are reconstructed and skimmed events. These come either from the actual real data or from the generation of simulated events (in jargon Monte Carlo, often misspelled “Montecarlo”, or MC).

  • The skimming is necessary to reduce the size of the dataset. This will simplify and speed-up the analysis.

  • The reconstruction step is the same for both real and simulated data to minimize the differences between the two, except that the data need to be first calibrated.

The following sections will quickly go through each of the four blocks in which the workflow is split, covering the very basic concepts and deferring most of the technical explanations about how the software works to the other chapters.

Author(s) of this lesson

Umberto Tamponi, Martin Ritter, Oskar Hartbrich, Michael Eliachevitch, Sam Cunliffe