Skip to content

Comparing and contrasting datasets

Often people are quite confused about the differences between flagging values in SDSS data. Here's how you can get a quick look at it.

Example: looking at the impact of flagging columns#

Let's create three subsets. Both on the ASPCAP dataset, but with three differences in flagging.

  1. No filtering at all
  2. flag_bad == 0
  3. result_flags == 0

Note

You must disable the default purely non flagged option in the Quick Flags, under the Targeting Filters menu.

setup

The setup described above. Use the Expression field to setup the flagging checks.

Now create 3 heatmaps, plotting teff vs logg and set the bintype to count with a logscale. Make sure to flip the axes, and set the Subset in each to the correct one.

plots

The setup described above. Top left: no filtering. Top right: flag bad == 0. Bottom: result_flags == 0. Look at the structural differences.