Galaxy Zoo Starburst Talk

The Galaxy Zoo forum Object of the Day, for Saturday 31st August, 2013

  • JeanTate by JeanTate

    It's a follow-on from the one I wrote two weeks' ago, and is called A Paler Shade of White.

    Here's my attempt to reproduce it, in Talk format (I don't know how to remove the white space between the rows of images)

    A couple of weeks ago I wrote an OOTD called simply White.

    In it I had a scatterplot of the colors of the Quench Sample galaxies, from the Quench project; here it is again:

    In this OOTD I'm going to take a closer look at the main body. Here's what I did:

    • combined the Quench Sample and Quench Control
    • selected only those smaller than ~1.1" (as determined by the value of
      their Petro_R50 parameter)
    • removed various outliers (e.g. those obviously much bigger than
      1.1"      )
    • transformed the central part (see below) into a 10 x 10 grid
    • selected a representative galaxy for each cell, if there was one;
      otherwise, left the cell blank.

    Here's a plot of all galaxies, after selecting the small ones and removing the outliers. The 'central part' is defined by the orange lines:

    So here's that central part, populated with the DR7 images of the Quench project galaxies:



















    In a later post in this thread I'll provide a list of DR7 ObjIds for each posted galaxy. Also - if anyone is interested - the detailed steps involved in the transformation (and selection).

    Perhaps the strange color tints in many of these images have more to do with the way the images of small galaxies are processed in DR7 than with intrinsic color gradients in the galaxies themselves. Judge for yourself ... here are the same objects, per DR10 (separate post, because otherwise I'll exceed the 20k character limit).

    Posted

  • JeanTate by JeanTate

    The 'paler shade of white' chart as seen through the eyes of DR10:



















    Posted

  • JeanTate by JeanTate

    Test, of an extension of the approach I'd used to create the 10x10 collage of small images. In this, I used, as input, a list of 100 (RA, Dec) coordinates, the scale parameter (0.4 in this case), the image size parameter (60 and 60, in this case; for some reason it gets implemented as 64x64), and the text strings for the pieces of the URL to get a DR10 image cutout.

    These are the Quench Control objects, sorted by Abs_r, the estimated absolute magnitude in the r band (a.k.a. Mr), from intrinsically faintest in the top-left corner, increasing to the right, then down.

    AGS00004d5AGS00002y7AGS0000352AGS00004azAGS00004igAGS00004ahAGS00004lbAGS00002udAGS00002p7AGS00004ak
    AGS00004eyAGS00004a5AGS00002dzAGS00002dfAGS0000445AGS00004bmAGS00003uyAGS0000383AGS00003idAGS00002g9
    AGS00002emAGS0000457AGS00004lyAGS00003gwAGS00004kqAGS00004lsAGS00002m6AGS000048aAGS00003clAGS000045d
    AGS000034rAGS00004k3AGS00002ldAGS00004m9AGS00004ksAGS00004avAGS000044fAGS000033zAGS00004c6AGS00002zd
    AGS00003meAGS00004mmAGS00004e5AGS00002p2AGS000049hAGS00002ooAGS00002o2AGS000047nAGS00004inAGS00002md
    AGS0000329AGS00002oxAGS00003tnAGS000047kAGS00004efAGS00002hnAGS00003q7AGS00003y9AGS00002ijAGS000041c
    AGS00004juAGS00003e5AGS00003zeAGS00004esAGS00003toAGS000037oAGS00002jpAGS000031pAGS00003qtAGS0000459
    AGS000048cAGS0000314AGS00003xaAGS000036zAGS0000418AGS0000434AGS00004hhAGS00004b3AGS000042sAGS000041g
    AGS00003mdAGS00004loAGS00004i7AGS00002k9AGS000048vAGS000044dAGS00004ilAGS00002gtAGS00002fbAGS00002c4
    AGS00003u3AGS00003w0AGS00003qpAGS00002wkAGS00002viAGS00003xgAGS00002kdAGS00002gvAGS00003ioAGS00003x9

    It's pretty easy to pick out likely outliers - objects whose Mr are badly wrong - and to see that there seem to be an awful lot of highly inclined to edge-on spirals (and extremely elongated ETGs) ... though this impression may be wrong (but it triggers the thought to go investigate!).

    UPDATE: I replaced the image URLs with ones containing the AGS IDs, in the alt field.

    Posted

  • jules by jules moderator

    Wow - that's a lot of work there Jean! I too was struck by the number of edge-on / near edge-on galaxies in the above QC list as I was scrolling through - then you mentioned it! Interesting.

    Posted

  • JeanTate by JeanTate in response to jules's comment.

    Thanks jules.

    Turns out it's not that much work ... once the framework was worked out, tested, and tweaked until it worked (yes, that took some time). I applied this to the smallest 100 QC objects, with two different scales, here.

    UPDATE: There are, indeed, far more cigar-shaped galaxies in this 'first 100' than you'd expect, if they were distributed randomly among all QC objects. There are 17 in this 'first 100', and 158 in the whole QC catalog (3002 objects in total). On a standard X2 contingency table test, this has a p of 0. The corresponding numbers of 'edge-on' (Eos) are 9, 177, and 0.18; i.e. Eos are more common among the first 100, but not particularly (statistically) significantly so.

    I wonder, what about the QS catalog?

    Posted

  • JeanTate by JeanTate

    Here's the same thing, for QS (faintest 100, per the Abs_r field; smallest at top left, increasing to the right then down; DR10 images, scale=0.4):

    AGS00001z3AGS00002akAGS00000toAGS0000038AGS00000mtAGS000021rAGS00001zlAGS000029xAGS00001z6AGS00001yr
    AGS000023kAGS000016zAGS00000l6AGS000002lAGS000015iAGS0000021AGS00001wwAGS00001tzAGS000020sAGS00001jk
    AGS00000asAGS00001ttAGS00000azAGS00000tdAGS00000nzAGS00001srAGS000008bAGS00000j6AGS00000wpAGS0000231
    AGS00000slAGS00000dtAGS00001t1AGS000009zAGS000029cAGS00001w6AGS00000doAGS000022rAGS000023eAGS00000iz
    AGS00000qbAGS000029eAGS00001etAGS00002avAGS00002aeAGS0000069AGS00001e0AGS00001tvAGS00000s7AGS00000co
    AGS000028gAGS00001blAGS00000pqAGS00001y3AGS00000tzAGS00000otAGS000007gAGS00001zpAGS00001n5AGS00001ia
    AGS000017aAGS0000238AGS0000208AGS00000qvAGS000012rAGS0000272AGS00001pwAGS000026tAGS000008zAGS00001gr
    AGS00000cnAGS0000117AGS000009cAGS00001y4AGS00000daAGS00001azAGS00000w9AGS00000d4AGS000008vAGS00001zh
    AGS00001kaAGS00001mvAGS000023dAGS00001szAGS00001u9AGS0000279AGS00000bzAGS00000dmAGS00001omAGS000027i
    AGS00000ezAGS00000waAGS00001hvAGS00001w9AGS0000055AGS00001y9AGS00001o0AGS00001a3AGS0000190AGS000008m

    Again, there seem to be rather a lot of Eos and cigar-shaped galaxies, as well a quite a few "mis-identified" ones.

    UPDATE: the numbers are:

    • 15 Eos in the 'first 100', 231 in the whole QS catalog
    • 18 and 224 cigar-shaped, respectively

    A difference - higher incidence of both Eos and cigar-shaped among the first 100 - which is (quite) statistically significant. Oh, and I should have noted that I made no attempt to remove, or otherwise account for, the known duplicates (in both catalogs).

    Posted