Python · SQL · scientific data quality

SDSS Data Quality and Survey Completeness

A reproducible analysis of survey coverage, missing spectra, filtering logic, and large-scale galaxy data.

164,072galaxy records
85.918%with spectra
62,8793D-mapped sample
1,306filtered nearby objects
Bar showing 140,967 galaxy records with spectra and 23,105 without spectra out of 164,072 total.

What I did

I wrote SQL filters for clean galaxy samples, corrected observed magnitudes for extinction, used Python masks to separate records with and without spectra, and kept the sample counts visible at each step. A second analysis transformed sky coordinates and redshift into Cartesian coordinates for spatial visualization.

Quality mindset

Some later plots produced patterns that did not match the expected physical interpretation. Those outputs were flagged for calculation review rather than presented as settled findings. The public portfolio includes the verified portions of the work and leaves out unfinished Monte Carlo cells.

Project files