SDSS Photometric Measurements with Labels

The Sloan Digital Sky Survey (SDSS) is a comprehensive survey of the northern sky. This dataset contains a subset of this survey, namely the photometric measurements and spectroscopic labels of around 2.8 million objects. The dataset was generated by submitting an SQL query to the DR12 catalog on the SDSS CasJobs site.

Each row in the dataset corresponds to an object in the sky. There are 14 columns:

• The first two columns contain the right ascension (ra) and declination (dec) of an object. These two coordinates uniquely determine its position.
• The third column (class) is the spectroscopic class (Star, Galaxy, and Quasar) as determined by expert opinion. This can be the target vector in a classification model.
• There are 11 columns that we can use as feature vectors. These are the different PSF and Petrosian magnitude and colour measurements:
  • psfMag_r_w14: The PSF magnitude measurement in r-band, assuming the object is a point source.
  • psf_u_g_14: The difference between the PSF magnitude measurement in u-band and the g-band (i.e. the u-g colour), assuming the object is a point source.
  • psf_g_r_14: The difference between the PSF magnitude measurement in g-band and the r-band (i.e. the g-r colour), assuming the object is a point source.
  • psf_r_i_14: The difference between the PSF magnitude measurement in r-band and the i-band (i.e. the r-i colour), assuming the object is a point source.
  • psf_i_z_14: The difference between the PSF magnitude measurement in i-band and the z-band (i.e. the i-z colour), assuming the object is a point source.
  • petroMag_r_w14: The Petrosian magnitude measurement in r-band, assuming the object is an extended source.
  • petro_u_g_w14: The difference between the Petrosian magnitude measurement in the u-band and the g-band (i.e. the u-g colour), assuming the object is an extended source.
  • petro_g_r_w14: The difference between the Petrosian magnitude measurement in the g-band and the r-band (i.e. the g-r colour), assuming the object is an extended source.
  • petro_r_i_w14: The difference between the Petrosian magnitude measurement in the r-band and the i-band (i.e. the r-i colour), assuming the object is an extended source.
  • petro_i_z_w14: The difference between the Petrosian magnitude measurement in the i-band and the z-band (i.e. the i-z colour), assuming the object is an extended source.
  • petroRad_r: The size measurement of the object in r-band in arc seconds.

The measurements have been corrected for dust extinction (the scattering of light by the galactic dust) using the correction set provided by Wolf (2014). Each feature has also been standardised to have zero mean and unit variance.

For the code to generate this dataset, please go to the Github repo: https://github.com/chengsoonong/mclass-sky

If you use the SDSS data in your papers, please see here for instructions on how to cite: http://www.sdss.org/collaboration/citing-sdss/

Please also cite this upload if you have used this particular pre-processed dataset.