Quality Assurance in Final OPS
Processing
Overview:
o All data have been through detailed QA during preliminary data processing.
o Only data known to meet requirement for level-1 specs are being run, so
emphasis is on testing for changes relative to preliminary processing.
o Operational efficiency is essential. QA cannot be a limiting element of
the data processing system.
The basis of our quality reviews is to:
(1) determine the extent of score/diagnostics changes when compared to
results of earlier processing.
(2) perform checks on functionalities new to the final processing pipeline.
(3) add new science diagnostics to check validity of results against an external
reference Despite the extent of data checking performed during final processing, the
quality assurance system has been built for speed and efficiency. Most
checks are automated so that potential problems are flagged for further
investigation by the reviewer. Diagnostics are, where possible, also shown as
plots for easy and quick interpretation.
QA review template
o Provides recipe book for reviewing a night.
o Makes sure reviewer has consulted all plots and other diagnostics for
potential problems.
o Current view of the plots and diagnostic tables used for QA.
o Are the locations of sources on color-mag and color-color plots (high latitude and low latitude) reasonable?
o Do the colors of saturated Read-1 sources fall in reasonable locations on color-color diagrams?
o Do plots of the photometric uncertainty as function of magnitude meet expectations?
o Are there any biases in the difference between PSF and aperture photometry as a function of magnitude or cross-scan position?
o Do plots of average color vs. cross-scan position show any biases?
