Galaxy Zoo Talk

GZ Survey numbers and other surveys

  • Peter_Dzwig by Peter_Dzwig

    Trawling through GZ data from Hubble I noted that some at least of the objects that had Survey Numbers and Survey Names attached were not actually listed in the relevant survey under that number. This question is firstly in relation to that issue, but I guess is more broadly applicable

    Three questions arise:

    (i) what do the Survey Numbers refer to (ie which surveys)? As an example, for one particular object the GZ ID was linked to AEGIS, but with a reference number that actually referred to a GALEX ( I think) survey number. Moreover some of the data given by AEGIS did not agree with that shown on the GZ examine page. This included RA and Dec, although the NED page was supposed to be centred on the RA and Dec given by the Zoo.

    (ii) What do the AHZ IDs actually refer to? Is there a link between them and other surveys?

    (iii) How is the GZ quoted data (on for example the examine page) arrived at? It appears that it doesn't always come from the other survey data. I have for example found redshifts varying by as much as z=1.5 on occasion, which isn't just photo z vs the rest when the actual survey data referred to is maybe 0.4 in one place and 1.9 in another

    Can anyone clarify?

    Thanks

    Posted

  • vrooje by vrooje admin, scientist

    Hi Peter,

    I'm not exactly sure what survey ID you are referring to, but it might be that the paper and catalog by Griffith et al. (2012) would be helpful. The original Galaxy Zoo: Hubble metadata was taken from an early version of that catalog, so the Hubble IDs on the Examine page for those old subjects should correspond to the IDs from that catalog. If I recall correctly they're integer IDs with a fixed number of digits, where the first digit tells you what survey the galaxy is from. Redshifts, magnitudes etc. were all from that catalog, but were in many cases assembled from other catalogs -- the paper would have more detail on those sources. I wouldn't be surprised if some of the redshifts had been updated since then with better data, sometimes much better data.

    The AHZ IDs are internal Zooniverse IDs for each subject, so they won't help you when looking through any external catalogs. We have a table that links them to other survey IDs, of course, but even in the absence of that you could use the RA and Dec.

    I should note as well that the latest Hubble data doesn't use the Griffith et al. ID; the Hubble IDs in those cases are IDs internal to the relevant science team, e.g., GOODS or CANDELS, and won't help you much in a search.

    In general a positional search is the most reliable thing unless you are certain that a unique ID from a catalog matches that from another catalog.

    Cheers,

    -Brooke

    Posted

  • Peter_Dzwig by Peter_Dzwig

    Brooke,

    Hi. To clarify. The Survey Numbers that I am looking at are the one on the examine pages which are 8 - 12 digits long I believe corresponding to the survey involved) AEGIS ones tend to be longer

    Thanks for the link to Griffith et al. I suspect that at least some of the data has been updated. There always were at least some for which the quoted redshifts didn't tie in to NED or whatever. Searching by RA, Dec is fine, but (in the case that I referred to above for example) there isn't an object at the site quoted - and I am trying to do this automatically if I can.

    Thanks,

    Peter

    Posted

  • vrooje by vrooje admin, scientist

    Hi Peter,

    If you download the Griffith et al. catalog and do a spot check on those coordinates and IDs with a sample across the GZ Examine pages, do they match? If so then you might be able to use the redshifts there just as a first pass, but you could also separately download the more updated spectroscopic redshift catalogs for the GZH fields (such as zCOSMOS, which may have released a better catalog) and use those where available.

    Last time I did this (which was a while ago) I did it in TOPCAT, which was brilliant for this sort of thing.

    Hope that helps,
    -Brooke

    Posted

  • Peter_Dzwig by Peter_Dzwig

    Brooke,

    thanks from darkest Guildford! I have looked at Griffith et al, it's very interesting and I am going to plough through it in more depth. I haven't downloaded the catalogue yet though.

    This is something that has been rumbling on for some while. The issue comes about because of (i) inconsistencies across datasets; (ii) lack of identical centring and co-ordinate systems, though I think that all RAs, Decs are now all J2000 in GZ and centring is better.

    The nub of it is that often when you look at the object nearest the centre you often find that the catalog reference is wrong, or at least different and so is the redshift. To compound matters I have come across cases where none of the objects really matched the quoted z. I did put out a plea for an explanation of how the quoted GZ redshifts are obtained and how they relate to the figures in, say, NED or AEGIS. I am afraid to have to say that I got no response, although I am aware that it is not simple.

    I have used TOPCAT too, though some years back, but found it rather clunky so wrote my own i/f to the examine pages and went from there. 😉

    Peter

    Posted

  • vrooje by vrooje admin, scientist

    Hi Peter,

    Yep, I agree entirely and I wish I had the magic recipe for getting source identifications to always match and redshifts from different catalogs to always agree! But this is just one of those things that plagues us; it's generally best to just pick a way forward for deciding what to use and what to reject and make sure you can explain it clearly. That's what I do. 😃

    Good luck!

    -Brooke

    Posted

  • Budgieye by Budgieye moderator

    Here is a list of surveys in Galaxy Zoo Forum from several years ago.

    Finding redshift and spectra for Hubble galaxies in various databases http://www.galaxyzooforum.org/index.php?topic=277683.0 Some of the instructions will be out of date. Hopefully I will update it in Talk soon.

    Is this what you wanted?

    For matching coordinates from two sources, I did a quick survey of differing coordinates of quasars (point sources which don't move), and decided that if the two coordinates were 0.02 " of each other, (like in the NED column "separation arc-min" ) they were the same object.

    Posted

  • Peter_Dzwig by Peter_Dzwig

    Ah the Forum! Repository of so much knowledge!

    I will have a trawl through that page and see what I can identify.

    I built some software some time back to analyse pages (particularly from GZH) and am trying to use it to identify and/or confirm things like redshifts in both that survey and the new data for objects that I called BBOs (caterpillars).

    When I try to cross-check the survey names they don't tally with what is in NED and I know from experience that trying to cross-correlate on the redshifts is a very tricky thing, because they don't correlate well with GZ redshifts. Many moons ago I tried to get someone to explain where the GZ redshifts come from and got little more than that it's what the pipeline spits out.

    As Brooke said you can look at the data and take a view, but that isn't easily programmed and gets extremely tedious to do manually above a few dozen objects and I am talking several hundred - and I am not ZkK 😉

    Even if you could, there is the ubiquitous "....." in NED which may mean "ditto" or may mean " I dunno". Another question then: what does "....." mean? 😉

    I thought that you could use NED as an SQL (or NoSQL) DB (is that right?), which is one of the things that I wanted to do, but If all that I have is RA, Dec I risk getting lots of false positives. I guess what I am looking for is a meta-table, a sort of concordance - which would correlate survey IDs. A sort of meta-NED. I don't suppose that anyone has produced one have they?

    And yes I had thought of doing the obvious, but I don't have access to that sort of compute resource anymore

    Any thoughts are welcome

    Hope to see you in Oxford,

    Peter

    Posted

  • Peter_Dzwig by Peter_Dzwig in response to vrooje's comment.

    Brooke,

    for the reasons that I am trying to do this see my response of a few seconds ago.

    Peter

    Posted

  • Budgieye by Budgieye moderator

    A special prism diffraction plate called a GRISM (GRating prISM) is placed over a whole section of image, which produces rainbow coloured streaks instead of stars and galaxies.. The streaks are analyzed for amounts of light along their length, which produces a spectrum. I am looking for a link, can't find it.

    Posted

  • Peter_Dzwig by Peter_Dzwig

    There's a GRISM image in Schawinski et al here

    Posted