Description of the Challenge

SDC2 is a source finding and source characterisation data challenge on a simulated SKA HI data product.

Summarised details of the simulated data can be found here, and a full description of the data and the challenge is available here.

Participating teams are able to access the full-size data cube on dedicated facilities provided by our computational resource partners. Each team has been matched to a computational resource from those available, and will deploy there their data analysis pipeline. Smaller versions of the full data product are publicly available for direct download to help with the initial data inspection and pipeline design.

Teams will undertake:

  1. Source finding, defined as the location in RA (degrees), Dec (degrees) and central frequency (Hz) of the dynamical centre of each source.

  2. Source property characterisation, defined as the recovery of the following properties:

  • Integrated line flux (Jy Hz): the total line flux integrated over the signal ∫F d ν

  • HI size (arcsec): the HI major axis diameter at 1 M ⊙ pc-2

  • Line width (km/s): the observed line width at 20 percent of its peak.

  • Position angle (degrees): the angle of the major axis of the receding side of the galaxy, measured anticlockwise from North.

  • Inclination angle (degrees): the angle between line-of-sight and a line normal to the plane of the galaxy.

Submissions must strictly adhere to a specific format, which is be detailed in the SDC2 Scoring Document and is demonstrated by the truth catalogues accompanying the small 'development' datasets.

An official scoring service is available, to give participants the means to self-evaluate their performance and improve on it. Credentials for use of the service are provided to participating teams. The service can be run from any machine and is available as a pip-installable package.

The total size of the HI challenge data cube is around 1 TB.