Radio Galaxy Zoo Talk

Seeing the ATLAS subjects more often

  • KWillett by KWillett scientist, admin, translator

    Hi there,

    I wanted to give everyone a quick update regarding the frequency of the different data sets on RGZ. The total data sets are only about 2% ATLAS and 98% FIRST images - mostly due to the fact that the FIRST survey was much larger and had more subjects. ATLAS was deeper, but only covered a small area of the sky (~7 square degrees). We'd really like to finish the ATLAS subjects soon, though, since there are only 2460 images and we could complete the full analysis of that dataset soon. That'll mean ATLAS will be finished completely for the first data release paper and allow us to do a direct comparison to existing papers with published results (which help us better calibrate the consensus algorithm).

    So the main result is that we've increased the number of ATLAS images active on RGZ; you should now see them at a 50/50 ratio with the FIRST images. The classification interface remains exactly the same, but with the increased frequency, we're going to shoot for finishing ATLAS within a couple of months.

    As always, thanks for your help and continuing to help us by classifying galaxies. We're excited to be able to start working on the finished ATLAS subjects soon. Let us know if you have questions!

    • Kyle

    Posted

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

    This is good news, Kyle, thanks! 😃

    Can you give us the center of the area on the sky for the ATLAS images (assuming it's a single, contiguous area)? I suspect that there's zero overlap with SDSS - so no point in clicking that link - and also ~none with either FIRST or NVSS. If so, this means that it would be quite challenging for us ordinary zooites to check any field, to see what the host might be, or try to run down any possible giants, etc.

    Posted

  • KWillett by KWillett scientist, admin, translator

    ATLAS surrounds the Chandra Deep Field South (CDF-S), in the Southern Hemisphere. The center is roughly at RA = 3h36m00s, Dec =-28d00m00s.

    Norris et al. (2006) is the main data release paper of the radio observations; you can see a plot of the whole field in Figure 1.

    Posted

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

    Thanks!

    Posted

  • raynorris by raynorris scientist

    That's great - many thanks Kyle. We're trying a whole load of different approaches on cross-ID'ing this field, as a learning experience in preparation for EMU ( http://www.atnf.csiro.au/people/Ray.Norris/emu/index.html ). As well as RGZ, I'm doing them by eye, a PhD student (Stuart Weston) is trying a likelihood-ratio approach, and we're also trying a full-bayesian approach (see http://adsabs.harvard.edu/abs/2015MNRAS.451.1299F ). We're also planning a machine-learning approach, using Amazon's cloud computing platform. EMU, with 70 million galaxies, will probably require all these approaches, and this experiment with RGZ will start teaching us which approaches work best in which types of source.

    And of course, as always with RGZ, we're hoping that RGZ may give us some surprises too!

    Ray

    Posted

  • 42jkb by 42jkb scientist, admin in response to KWillett's comment.

    ATLAS also images the ELAIS S1 (European Large Area ISO Survey South 1) region too.

    Article describing the data is here: http://adsabs.harvard.edu/abs/2015arXiv150803150F

    Posted

  • HAndernach by HAndernach scientist, translator

    Are the ATLAS subjects also 3' x 3' in size? For ATLAS images the links to SDSS and FIRST are indeed useless,
    while NVSS is often available, but MUCH LESS sensitive than ATLAS.
    It would be helpful to be able to display a larger image of ATLAS to recognize overedge sources,

    Posted

  • 42jkb by 42jkb scientist, admin in response to HAndernach's comment.

    The ATLAS images (2' x 2') should be linked to larger 5' x 5' fields of view.

    Posted