Galaxy Zoo Talk

useful or not Comments

  • liometopum by liometopum

    Since I am in a chatty mood...
    Many zooies like to list the galaxy catalog number in the comments, and some people list a bunch of code-looking information that I have no clue as to what they are saying.

    Is that stuff worthwhile to add to the Comments? is it of value to the professional users of our citizen classifications? The galaxy numbers should be associated with the file anyways, so is this a waster of time or is it desired?

    Posted

  • Budgieye by Budgieye moderator

    I don't understand some people's code information either. Sometimes it is their own private code, and moderators try to discourage that.

    Sometimes it is good to add other names to comments such as #NGC numbers . Sometimes there is more than one image of a galaxy. Also, different numbers can be used at different times eg. 58.............., 12................, AZH............... . Then we can find the galaxy under different names.

    Most of the information on Talk is for the fun and education of people who post. It is only by learning what is usual or already known, can we then recognize when something is unusual and should be reported to the scientists.

    Posted

  • Capella05 by Capella05 moderator

    Following on from what Budgie said, at lot of the 'code' is just data copied across from Skyserver / NED / SIMBAD so it is already part of database / dataset. The only purpose copying it across would be too inform your fellow classifiers, or to build up a collection.

    I prefer to limit the amount of hashtags I use, as I feel they can be off putting to people who are new to the site and don't understand them, but each to his own 😃

    Edited to add: the hashtags that we really want to see are the features (or anything unusual) that you noticed when classifying the image

    Posted

  • JeanTate by JeanTate

    Christian+ (2012) "Citizen Science: Contributions to Astronomy Research" is the only paper I know of which describes the professionals'/scientists' aims for Talk's Comments (etc)*; it's an interesting read.

    Despite whatever hopes (etc) Talk's prime movers may have had, for generating Real Scientific results from Talk's Comments, I don't think there's been anything concrete yet; in particular, as far as I know Zooniverse's Talk has not been even mentioned in any paper reporting Zooniverse-based crowd-sourcing/citizen science results, much less crediting anything in any Comment for any insight, discovery, etc (if anyone reading this knows of any concrete examples, please post them! 😃).

    Sometimes it is their own private code, and moderators try to discourage that.

    One of Talk's obviously great potentials is Collections+hashtags+Search. Especially for zooites who want to do their own research (no matter how serious). If it all worked as per the spec, it'd be wonderful! A consequence is: zooites will add their own 'codes', in Comments, as an aide to collecting/classifying/etc; especially if they're had frustrating experiences with the multiple failures of Collections+hashtags+Search to perform as expected.

    So if mods are discouraging this, without knowing what the zooite is trying to do, it may be sending very much the wrong message re scientific curiosity and exploration (IMHO).

    Sidenote: occasionally this can throw up strange things, such as when a zooite describes a morphological feature in a galaxy image as a 'lens' - with the standard deVaucourleurs+ meaning - but another mistakes this for 'gravitational lens'! 😛

    *Robert Simpson has a draft paper which presents some analyses of how Talk's Comments compare with Talk's Discissions (he provided a link to it in the GitHub discussion on the (now submitted) Marshall+ (2015) AURA paper on Citizen Science in Astronomy); as far as I know this has not yet been 'published' even in draft form

    Posted

  • DZM by DZM admin in response to JeanTate's comment.

    One of Talk's obviously great potentials is Collections+hashtags+Search ... especially if they're had frustrating experiences with the multiple failures of Collections+hashtags+Search to perform as expected.

    This is as frustrating to me as it is for everyone here, but I do want to make it clear that next-gen Talk development is progressing! Apparently we're very close to being able to hook up the back end to the front end, and after that it will be all about design and functionality and making everything as user-friendly as possible. It will be extensively tested and revised, and hopefully should bring out the true potential of collections, hashtags, and the search function... not to mention notifications, cross-project discussion, improved navigability, and a shared Zooniverse space/identity.

    I still have no timescale, other than that it is a mandatory deliverable by the end of this year. I want it up soon, but I want it up perfect!

    Posted

  • liometopum by liometopum

    As a zooite, I want to make the best classification I can, however...

    1. The best tools come after the initial classification, so my opinion often changes after the first pass.
    2. I had thought that the serious observations should be listed in Comments, again because all the tools are available. But then, I read that this section is more for fun and sharing.

    If it is the first section of classifiying that is the important part, it would be nice to be able to use the tools from the start. That is, let us get into the SkyServer Imaging immediately.

    Posted

  • Capella05 by Capella05 moderator

    There is a reason why you are not allowed to use tools from the start. We want your unbiased opinion on the object you are classifying 😃

    Not Skyserver / NED or SIMBAD's - we have that already. We want to know what you see, and how you classify it.

    Also, SkyServer is not always correct 😃

    Posted

  • JeanTate by JeanTate in response to Capella05's comment.

    Also, SkyServer is not always correct 😃

    It's actually more interesting (and complicated) than that!

    As I noted in this thread, even for quite 'local' galaxies, detailed investigation by the experts reveals that morphological classifications depend at least somewhat on the band(s) used; for example, 'blue' bands highlight regions of recent star-formation, while 'red' ones the locations of older stars, so spiral arms may be prominent in one but invisible in another! 😮 Bill Keel (NGC3314) made this point rather dramatically, in an old GZ forum thread (which I can't find just now 😦).

    Posted

  • liometopum by liometopum

    The part of SkyServer that i use is the imaging. I can see th galaxy from far away, and zoom in, and do this also with the invert. The image on SkyServer sometimes is more revealing than what is given when the object is first presented. Numerous times, after looking at the large SkyServer image, and zooming in and out, and onverting, I can find a 'sweet spot' that shows me something I could not see in the initial image. i don't need the data. But the imaging ability within SkyServer is a big help. I have spotted rings, and almost invisible bars and arms, and mergers, this way. The better imaging, without the data, would allow a better unbiased opinion.

    So maybe give classifiers the more versitile imaging ability.in the next go-around.

    And let me throw this in. I ask myself why I take the time to do galaxy classifications.

    1. To learn about galaxies, first-hand
    2. I like to think that the effort is of use to others; that I am contributing to something.
    3. To find something new. (How many times have people got excited thinking they found a new voorwerp?)

    For example, I have had some recent exchanges with capella05 about a possible merging galaxy. Outside of the intial classification, it appears no researcher would notice,the comments, and I am afraid all the fuss is for nothing. And this galaxy also is one that I had to use the large SkyServer image to really appreciate and (hopefully) correctly classify. It was not clear to me until after I had clicked on the link to Discuss This, and looked a length at the large SkyServer image.

    Some of the big guns (vtaskew, d Graham etc) should give their thoughts on what they/'d like to see in an upgrade.

    Posted

  • JeanTate by JeanTate

    It's been a while, so I may not be remembering everything correctly, but in the first GZ, zooites could zoom in and out, and invert, before they had to commit to a classification. But when it came to turning all the zooites' clicks into classifications, there was very little segmentation (as I recall); see the GZ papers which contain the details of how clicks were turned into classifications. In a way this is somewhat strange; there's a paper with Arfon Smith as an author - on the SupernovaZoo, to be sure - which shows that zooites can be quite cleanly segmented. And there's a GZ blog post by zkChris on how zooites' strengths (wrt classification) could be discovered and leveraged to improve the efficiency and effectiveness of such projects (there was also a week-long workshop - in Taiwan? - which explored this further).

    A suggestion, if I may: it's very hard (IMHO) to discuss classification(s) in just 140 characters; a Discussion thread is much better (despite all the shortcomings of these, in the current version of Talk). Just like this! You may also find this recent thread interesting: A Classical Morphological Analysis of Galaxies in the Spitzer Survey of Stellar Structure in Galaxies (S4G)

    Posted

  • liometopum by liometopum

    Another note of improvement: when we collect a galaxy, the group name goes to the top of the list, so the list keeps getting shuffled. I hate that.

    It makes me stop every time and read my list to find the collection group. It needlessly slows me down and wastes classification time. Placing the most recently used group on the top is not helpful, as the presentations are random.

    Better to either keep the same arrangement, or let us have the ability to resort the list as per individual preference, and leave it untouched by the classification process.

    Posted