search

dfm / Star-Field

Other
5 stars 1 forks source link

plot SEx and DOLPHOT more fairly #21

Closed davidwhogg closed 11 years ago

davidwhogg commented 11 years ago

On the plot where we compare SEx to the truth, we plot the truth as a horizonal line at 100 or 1000; really we should plot the cumulative number of sources as a function of brightness. We can also do that for SEx catalogs because they return for each source a brightness. Same for DOLPHOT when we have it. We should also plot a few lines for samples from our catalog sample. That would be more fair.

eggplantbren commented 11 years ago

I'm trying this at the moment, but I'm having a problem because I don't know the proportionality constant between my "flux" values and SExtractor's "sigmas". I tried aligning the curves at the right end but they don't actually match up super well.

davidwhogg commented 11 years ago

You don't have to plot against sigma -- you can plot the cumulative number-flux relationship using SEx's reported flux values, and do a few lines for different SEx thresholds. That is, I am suggesting changing the x axis to flux and plotting the catalogs as number-flux relations (which is what we are on about anyway).

eggplantbren commented 11 years ago

Makes sense. I shall do that. On Nov 30, 2012 2:37 PM, "David W. Hogg" notifications@github.com wrote:

You don't have to plot against sigma -- you can plot the cumulative number-flux relationship using SEx's reported flux values, and do a few lines for different SEx thresholds. That is, I am suggesting changing the x axis to flux and plotting the catalogs as number-flux relations (which is what we are on about anyway).

— Reply to this email directly or view it on GitHubhttps://github.com/dfm/Star-Field/issues/21#issuecomment-10874838.

eggplantbren commented 11 years ago

Done (for SExtractor anyway. DOLPHOT is beyond the scope).