Improving the efficiency of SDRAGN discovery, an open discussion

by JeanTate

How can ordinary zooites and professionals (RGZ science team members, and more) organize themselves to do 'SDRAGN research' more effectively?

Background: In RGZ, SDRAGNs ('spiral double radio active galactic nuclei') are one of the big, ordinary zooite-led, discoveries, along with giants, hybrids, and more. Ray Norris' post - OP of the Hourglass sources associated with spiral galaxies thread - 13 months' ago kicked this off. Many candidates have been posted, and some have been investigated.

Motivation: However, the hunt is not very open, nor particularly systematic; there are few goals (beyond the vague 'find candidates'), etc. And if I were to be run over by the proverbial bus tomorrow, much of the work would be lost (I've posted far less to Talk than I have in my offline databases, for example).

Action: I'd like to kick off a conversation on how we could do better. And that's what this thread is.

Goal? To have something concrete to start discussing, how about, as a goal, an at least provisional agreement on how to 'publish' what's been discovered (and researched) so far, re SDRAGN candidates (and to what extent that publication should be open, or not)?

One additional point of interest, for this as a goal, is that it might bring into the open some of the peculiarities of how astronomy-as-science is actually done, and also encourage greater involvement.

Thoughts?

Posted January 21, 2015 12:44 AM
by DZM Zooniverse Team

Thanks for starting this up, Jean! 😃

I will probably have less to contribute to this discussion than will people who better understand the process by which "hey, this looks interesting" becomes accepted science and/or a published paper, but nevertheless, I will stay abreast of the conversation!

One thing that I can suggest, Jean, is that you allow someone whom you trust access to your offline databases? I'm not sure if you've kept more work there because Talk wasn't fulfilling your needs for the work you wanted to do, or because you wanted to work on it privately? Either way, a good way to at least guard against losing all of your work would be to let in a collaborator or two, someone who understands how you've organized things? Just a thought...

Another would be to at least get an official SDRAGN tag group up here on Talk. I might suggest two official tag groups made by a scientist or moderator: a "candidate" one and a "strong candidate" one. The "candidate" one would collect everything with an #SDRAGN tag; the "strong candidate" one would collect only images with #SDRAGN and a confirmation tag, such as #SDRAGNstrong , or whatever is agreed upon. That would allow experienced SDRAGN hunters to parse down anything that has even been mentioned as a potential SDRAGN to the items that really need additional research. (This is a strategy we're going to use with secondary classification for an upcoming project.)

On the topic of additional research, I would be interested in knowing at which point, if any, in this process, it's necessary for Talk folks to turn things over to the people with the big telescopes and fancy equipment. Is there a limit to how much even our savviest volunteers can do as far as SDRAGN identification/research without pro equipment?

Just my first few thoughts/ideas, for whatever they're worth. I'm always happy to see more of our work being down out in the open and with more collaboration!

Posted January 21, 2015 7:08 PM
by DocR scientist

Ray's input needed here.

Posted January 22, 2015 2:15 AM
by JeanTate

Thanks DZM, DocR.

As even casual readers of RGZ Talk will surely know, I am an enthusiastic zooite, keen as mustard on astronomy, and was over the moon when I discovered this astronomy Zooniverse project (here's a secret: I used to think radio astronomy boring, compared with optical astronomy; I could not have been wronger!). And if you're an oldbie zooite, who used to hang out in the Galaxy Zoo forum - especially the Object of the Day board - you'll know that I like talking about astronomy-as-science, Open Science, and so on.

So this thread is me trying to get a discussion going around these big, 'meta', subjects, by choosing a particular research topic, one that's 'hot', and one that I've been deeply involved with (and one that many other zooites have too).

Enough with the background, time to discuss astronomy!

At the level of "let's get started on a paper", I see two extremes: a paper on a particular galaxy/double lobe radio source - like J1649+26: A Grand-Design Spiral with a Large Double-Lobed Radio Source - Mao et al. - or a paper on "101 candidate SDRAGNs discovered in Radio Galaxy Zoo" (sorta thing).

What would each kind of paper look like? Well, let's discuss that! 😃 How would/could we go about choosing an object, or "101" objects, for each such a paper? Let's discuss that too! 😃

About Open Science: if we - collectively, somehow - decide we want to do the work and write a paper (one that's at least good enough to get up in arXiv (the astro-ph section)), how do we do this? Here, in RGZ Talk? Or perhaps we would be far too presumptuous, ambitious, (worse?) to think that we could? Let's discuss! 😃

And there are practical matters ... we can't dump 1MB of data into an RGZ Talk Discussion post, for example (not practically anyway), and if we use something like Dropbox or Google Drive (what's your fave?), what format do we use for the data? Let's discuss! 😃

Too, is RGZ Talk a suitable platform to use for collaborating on an Open Science research project? for writing a paper? Let's discuss! 😃

... you get the idea, right? ...

Now the above is all rather general; to close, something quite concrete, specific: the "S" in SDRAGN, or, "How do you know you it's a spiral when you catch a galaxy?" Turns out this isn't as straight-forward as it seems 😮 Let's discuss!! 😄

Posted January 22, 2015 3:21 PM
by leonie_van_vliet

Hi Jean,
Just read the above discussion. And oh my... definitely impressed. So many opinions and ideas exiting though as you see how
many people react. Encouraging and for now leaving this to you with the knowledge that you started something worth while.
Already learned something: Talk and share... 😃

Posted January 22, 2015 4:22 PM
by mini.mintaka scientist

Hi Jean et al 😃

Thank you SO much for starting this conversation!

As you know, i am very excited about spiral DRAGNs! I think they contravene the standard galaxy formation models because to trigger large-scale lobes/jets you need a major merger (e.g. Chiaberge & Marconi 2011) ... but spiral structure in a galaxy cannot typically withstand a major merger ... (shameless self-promotion... http://arxiv.org/abs/1410.8520)

The 'find more S-DRAGNs' project launched because we are interested in how prevalent these sources are... are there heaps of these things that have been somehow overlooked?? or are they very very rare and require specific conditions to exist?

I think the hard work of you (Jean) and many others have identified some excellent S-DRAGN candidates, but I also think it's safe to say that there is no glaring example of a 'bog-standard' spiral galaxy hosting large-scale radio lobes... what does this mean?? that something special is required for these unusual sources to exist!

Moving forward I think it'd be great to quantify what we already know. Specifically:
- Would the four currently known spiral DRAGNs have been identified by RGZ-ites? I think the answer is yes - J1649+2635 is (i think?) the only spiral DRAGN in the current RGZ footprint and I think you identified that one ...
- What is the number density of the SDRAGN candidates? This is (tricky and) interesting because while (i think) none of the SDRAGN candidates are 'normal' spirals... many of these sources are very odd!! The first step is of course to weed out the chance-alignments (and i think in Jean's previous thread on the 'best' candidates we may have already done this...)
- What are the properties of the final sample of SDRAGN candidates? Are there global similarities??
This post is getting long so I'll leave it for now...

more later

clear skies

minnie

Posted January 22, 2015 10:30 PM
by Ptd

My knowledge of astronomy is limited to what I've learned in the Zoo but my thoughts on this are:
I think I read somewhere that the primary purpose of RGZ is to cross match the WISE and FIRST surveys and identify sources that appear in both. With all other discoveries made on top (no doubt expected to be loads 😃) being bonuses.

But we've been given links to two other surveys as well, SDSS(I know roughly what that is from GZ) and NVSS(don't know anything about this one and the images look all blurry). So its now my standard practice to see if I can identify a SDSS object that goes with the ones in FIRST and WISE and I highlight this in a comment each time (if for no other reason than I like looking at SDSS images). But I've never been sure if that has any real scientific worth or if either WISE or FIRST have already been matched against SDSS?

So that's the first question I'd ask if I was hunting SDRAGNs.
1. Is there a list somewhere of all the surveys that have been done either whole sky or partial?
2. What cross matching of these surveys has already been done?
3. Of those which ones are available to download off the Internet?
4. Which ones has someone already done a spirals hunt in?
5. Of 4. is that data available and how searchable is it?
From the above draw up a list of currently known identified/suspected spirals. This list will I'm sure grow with time.

Next job, see if its possible to use the RGZ interface to examine those particular patches of sky for hour glasses, triples etc. Rather than look at all those images personally maybe find a way of searching for relevant hash tags associated with the patch of sky you are interested in in case someone else has already done that job for you (that's what this place is for after all 😃). I'm assuming Talk uses a database of some kind, is there a column in that database which records the declination etc. whenever a hash tag is made? If so you could probably paint the tags onto the sky so to speak. Then you run a search for likely looking tags, within, a chosen number of arcsecs. of your spirals. etc....

Err sorry is someone has already done all that or if its gobbledygook 😦

Posted January 24, 2015 12:47 AM
by mini.mintaka scientist in response to Ptd's comment.

Hi Ptd!

Excellent points! The search for SDRAGNs is only now really starting to be possible on a large-scale-non-serendipitous-way because of optical surveys such as SDSS that have the resolution for us to 'see' spiral galaxies !! To search for SDRAGNs really what we want are deep optical surveys with the resolution to ID spirals (yay SDSS!) and radio surveys that cover the same regions (NVSS/FIRST/SUMSS etc...) All of these data are available online!

So when Ryan Duffin searched FIRST for SDRAGNs he used the galaxyzoo classifications as his first stop to identify all the sources in SDSS that had been classified as spirals. (we chose to use the superclean sample)... he then crossmatched this list with sources that we determined to be 'extended'. Extended sources were determined by cross-matching NVSS (1.4GHz VLA All-Sky Survey at D-array) with FIRST (1.4GHz 10000sq deg survey with B array). Basically NVSS is lower resolution, which means there is a higher sensitivity to low surface brightness objects. (I am quite short sighted so when i want to find the LMC/SMC in the sky often i take off my glasses so I have a 'lower resolution' view, and then i find them more easily 😃)

ANYhow... because the NVSS and FIRST are taken at the same frequency, cross-matches sources that have more signal at NVSS than FIRST are likely to be 'extended'. It is important that we filter out the 'compact' radio sources because many many spirals have compact radio emission, either due to star-formation, or a Seyfert-like AGN in its core...

I hope that sorta answers some of your questions. Please let me know where i can be clearer!

clear skies

minnie

Posted January 27, 2015 12:11 AM
by JeanTate in response to JeanTate's comment.

Thanks leonie, minnie, Ptd! 😃

I am very glad discussion in this thread is moving along nicely.

[me] Now the above is all rather general; to close, something quite concrete, specific: the "S" in SDRAGN, or, "How do you know you it's a spiral when you catch a galaxy?" Turns out this isn't as straight-forward as it seems 😮 Let's discuss!! 😄
.

[minnie] Ryan Duffin searched FIRST for SDRAGNs he used the galaxyzoo classifications as his first stop to identify all the sources in SDSS that had been classified as spirals. (we chose to use the superclean sample)

Punchline: small disk galaxies can be recognized as such, in SDSS images, only out to very small redshift; really ginormous ones, out to almost 0.4 (and perhaps beyond). In trying to decide if the obvious host of a doublelobe RGZ object is a disk galaxy, how do you take account of this (huge) selection effect, this bias?

Some details: Back on 16th May, 2012, over in the old Galaxy Zoo forum, I wrote an Object of the Day called "Seeing Is Believing (Or Is It?)". Here are six images from the OP that make my point rather well (refer to the OOTD for details):

These are all spiral galaxies. In fact, the first and third (in the top row) are very famous, they're M81 and M104 (the Sombrero). 😮

Here's one possible way to cut through this (severe) selection bias: look for Eos. What's an Eos? an edge-on spiral! 😛

A galaxy that has 'pointy ends' cannot be an elliptical; the pointy ends are the galaxy's disk seen edge-on (or almost edge-on). For example, z_sp 0.365 SDSS J091858.12-005303.1:

Of course, this is still a ginormous galaxy, far bigger than our own Milky Way, M31 (Andromeda), M104 (Sombrero), ... In fact, it's also bigger than almost all elliptical galaxies too.

But is this giant, distant Eos an SDRAGN? Sadly, no (but I'll do an overlay, just to be sure):

Posted January 27, 2015 2:34 AM
by JeanTate in response to Ptd's comment.

Minnie has already responded to Ptd's excellent questions and comments; I'm going to have a go at answering the five questions directly:

So that's the first question I'd ask if I was hunting SDRAGNs. 1. Is there a list somewhere of all the surveys that have been done either whole sky or partial? 2. What cross matching of these surveys has already been done? 3. Of those which ones are available to download off the Internet? 4. Which ones has someone already done a spirals hunt in? 5. Of 4. is that data available and how searchable is it?

Q1. Is there a list somewhere of all the surveys that have been done either whole sky or partial? There are many such lists; perhaps one of the most comprehensive is VizieR ... but because it's so comprehensive (it includes just about every astronomical survey done, ever!), it can be hard to sort it down to what's best for your particular purpose. For radio surveys, a really good combined catalog is the Kimball&Ivezić Unified Catalog (check out this RGZ Talk thread for details). For sky coverage that's ~the same as SDSS', the combo of FIRST and NVSS is hard to beat.

Q2. What cross matching of these surveys has already been done? Close position matching is something computers are very good at, and has been done, for just about every big astronomical survey. The Last of FIRST: The Final Catalog and Source Identifications is a thread - in Journal Club - on a recent, very detailed, matching (yes, it includes FIRST against SDSS)

Q3. Of those which ones are available to download off the Internet? In general, I really do not know, but would guess 'most'. For us, the Kimball&Ivezić Unified Catalog is perhaps the most relevant.

Q4. Which ones has someone already done a spirals hunt in? In one form or another, astronomers have been trying to find which spirals are radio sources since just about the beginning of radio astronomy, over a half century ago now. What they have found so far may be summarized thus:
- almost no spiral is an intrinsically strong radio source
- many spirals have 'point source' radio sources in their nuclei; many (most?) of these are AGNs
- many spirals have ~point source radio sources in their disks, sometimes several; many (most?) are supernova remnants and/or regions of intense star formation
- a small number of spirals have (relatively weak) radio jets, associated with radio-active (but weak) nuclear emission; however, in all but a very few cases these jets do not extend beyond the galaxy
- almost no spiral is the host of a classic Double Lobe (DL), or Hourglass (or FRII or hybrid or plume or ...), i.e. extended radio emission
Q5. Of 4. is that data available and how searchable is it? Roughly speaking, all of it is available, and all of it is searchable.

From the above draw up a list of currently known identified/suspected spirals. This list will I'm sure grow with time.

For SDRAGNs, the challenge isn't so much 'find spirals which have a radio source within their optical boundaries', rather it's to find spirals which are hosts of (detached, giant, etc) DLs (the 'DR' in the name). In several decades of searching, astronomers have found only ~five (the exact number depends on the exact definitions)*.

I hope this helps, and that an astronomer will come along and correct any mistakes I may have made.

*here is one, SDSS J140948.85-030232.5 ('Speca') (from p9 of the Hourglass sources associated with spiral galaxies thread; details there):

Posted January 28, 2015 3:49 PM
by JeanTate in response to JeanTate's comment.

How would/could we go about choosing an object, or "101" objects, for each such a paper? Let's discuss that too! 😃

Somewhere, well before choosing an object/"101" objects to write about but well after learning how to do RGZ classifications, I think we - the zooites actively participating in an Open Science project on SDRAGNs - would need to spend some time going over defintions and criteria.

Such as?

Such as what, exactly, do we consider a DR to be?

Such as how do we decide if a DR host is an S?

An extract from my post at the bottom of p4 of the Hourglass sources associated with spiral galaxies thread:
At this point, definitions and criteria become crucial.

For example, how do we count lenticulars (early-type galaxies with disks)? galaxies with 'disturbed morphologies' (e.g. mergers)? what criteria do we use - other than subjective opinions - to eliminate 'chance overlaps/alignments'? On the radio side, must there be unambiguous evidence of two lobes/jets (or will just one - extending beyond the optical boundary - do)?

In a nutshell, I think we should aim to find "any galaxy - other than a boring elliptical - which seems to have radio emission outside the optical boundary, or extending beyond that boundary"

How to decide if a likely host galaxy is 'a boring elliptical' or not? The usual zooite method (but blue ellipticals as hosts are definitely to be noted! they likely contain gas and dust), plus astronomers' tests (e.g. concentration ratio, color, Sérsic profile index).

So, a simple two-part test; how likely is it that:
1. this galaxy is the host?
2. this host galaxy is not a boring elliptical?
Scattered throughout the next ~dozen pages of that thread are posts in which I refine my selection criteria, and my criteria for my five ratings (excellent, good, fair, poor, not). In a while I'll edit this post to add what they currently are; perhaps we can then discuss them? 😃 And in a later post I'll explain (mostly summarize) how I collected candidates.

(It's now later)

Simply put, I try to answer those two questions, independently, on a five-point scale:
- Yes!
- yes
- maybe
- can't-tell
- no/not (for the 'boring elliptical' question, this means "it's certaintly a boring elliptical!")
There are two refinements in the 'galaxy is host' question:
- if the 'associated radio emission' does not extend beyond the optical (SDSS) boundary (or is not entirely beyond it, in the case of detached double lobes, say), then the rating is no/not
- I also ask 'how likely is it that there's radio emission from the nucleus?'
How do I rate a candidate, given the two (plus one) answers? Like this:
- excellent: Yes! and Yes! if also Yes! for 'nuclear emission', then it's a super-excellent 😛
- good: any combo of Yes!/yes and Yes!/yes (but not excellent; i.e. Yes!+yes, yes+Yes!, yes+yes)
- fair: not excellent or good, no 'can't-tell' or 'no/not' answers (other than for 'nuclear emission')
- poor: not excellent, good, or fair; no 'no/not' answers (other than for 'nuclear emission')
- no/not: at least one 'no/not' answer
- a Yes! or yes for 'nuclear emission' makes it easier to answer the 'this galaxy is the host' question; a can't-tell or no/not makes it harder.
To reduce subjectivity (which is inevitable), I first produce FIRST contour overlay images (on SDSS) - sometimes adding NVSS contours - centered on the candidate host, and do the rating in batches (preferably of at least 20). I have also done several 'sanity checks' where other zooites have - independently - provided me with their assessments.

(I have also done some quantitatively photometric analyses on a small subset of candidates, to see whether the candidate hosts are more like ellipticals or disk galaxies. This is not part of the above).

Questions? Comments?

Posted January 28, 2015 4:20 PM
by JeanTate in response to DZM's comment.

Today, in this post, I'll discuss some aspects of a comment or two by DZM, and how it relates to what I see as the broader 'Open Science, here in RGZ Talk' context:

One thing that I can suggest, Jean, is that you allow someone whom you trust access to your offline databases? I'm not sure if you've kept more work there because Talk wasn't fulfilling your needs for the work you wanted to do, or because you wanted to work on it privately? Either way, a good way to at least guard against losing all of your work would be to let in a collaborator or two, someone who understands how you've organized things? Just a thought...

This is good advice, DZM, and something I've started working on.

If we agree to do an Open Science project, one with the explicit, relatively near-term aim of writing a paper good enough to be submitted to the likes of MNRAS, do you think my databases should be made public? If so, how?

Another would be to at least get an official SDRAGN tag group up here on Talk. I might suggest two official tag groups made by a scientist or moderator: a "candidate" one and a "strong candidate" one. The "candidate" one would collect everything with an #SDRAGN tag; the "strong candidate" one would collect only images with #SDRAGN and a confirmation tag, such as #SDRAGNstrong , or whatever is agreed upon. That would allow experienced SDRAGN hunters to parse down anything that has even been mentioned as a potential SDRAGN to the items that really need additional research. (This is a strategy we're going to use with secondary classification for an upcoming project.)

That would surely be better than what we have today, in terms of tracking candidate SDRAGNs! 😃 However, it's got a few shortcomings, e.g.:
- as we know, Talk cannot find all instances of a particular hashtag, with 100% reliability (so a diligent researcher would still have to keep their own record)
- people mistype, even zooites! Even if Talk could find all instances of #SDRAGNstrong, it cannot find those with typos (e.g. #SRAGNstong)
- diligently putting all candidates into a dedicated Collection (or several) should increase the reliability; however, as we'd want the funnel to be wide at the top (capture even very poor quality candidates, lots of Type 1 errors*, to be sure we don't miss any, minimal Type 2 errors*), the SDRAGN Collection would quickly exceed 500 (and so be useless from then on)
- the metadata on candidates is very important, but cannot be cleanly and comprehensively captured using hashtags, Collections, or whatever; examples:
- who first 'discovered' it? you need to check both Comments and Discussions to find out (and if it's in a many-page Discussion thread, good luck with that)
- among all the Comments and text within Discussion posts, you'd want to extract the meat, on each candidate; how, copy/paste to a new thread?
- the elemental 'object' in RGZ is the field, which has a unique ARG000nxxx ID (n: digit, x: 0-9/a-z character); SDRAGN candidates are galaxies, plus FIRST/NVSS sources ... the mapping many-to-many
Here's a record in my main 'tracking' database (actually a tab in a spreadsheet)**,
```
22 2whf	ARG0002whf	SDSS J080259.73+115709.7 WizardHowl	March 26 2014 4:10 PM 0	0	0	-1	0	0	excellent\ WH top10/
```
The first is my (internal) reference/record number.

The next five are pretty obvious; the next six are, respectively:
- Is this the host? (0=Yes!)
- Not a boring elliptical? (0=Not a boring elliptical!)
- Radio emission from host? (0=Yes!)
- numerical rating (-1 = excellent; nuclear radio emission)
- Have I produced a host-centered FIRST contours overlaid on SDSS image? (0=yes)
- Have I posted this image? (0=yes)
The last field (column) is "comments"; pretty boring in the case, except that "WH top 10" means that it's one of WizardHowl's Top Ten.

In a separate tab there is data I need to create overlay images; as I may create more than one overlay per ARG field, or SDSS object, I keep these in a separate tab (if I were using a good database, this wouldn't be necessary, of course). Here's part of the record for the above SDRAGN candidate:
```
22 2whf ARG0002whf SDSS J080259.73+115709.7 120.7489 11.9527 0.16 0.132±0.0266/0.141±0.0327 0.18 FIRST 1 18 0.03 http://skyserver.sdss3.org/public/en/tools/explore/summary.aspx?id=0x112d13d8402c0268&spec=&apid=
```
How well (robustly, comprehensively, easily accessed and used, reliably, ...) does anyone think today's RGZ Talk could 'keep' such data?

*or is it the other way roud? I can never keep the distinction straight! 😦

**actually there are many more fields (columns), but these are the essential ones

Posted January 29, 2015 9:34 PM
by DZM Zooniverse Team

as we know, Talk cannot find all instances of a particular hashtag, with 100% reliability (so a diligent researcher would still have to keep their own record)

I will say this: the tag group feature actually does pull with 100% accuracy. It is limited to 500 images, but do you really expect over 500 candidates even at your first screening level? SDRAGNS are super rare, right?

people mistype, even zooites! Even if Talk could find all instances of #SDRAGNstrong, it cannot find those with typos (e.g. #SRAGNstong)

This is the problem with the tag groups only collecting "and" images, and not "or" images. Of course, there'd be no way to anticipate every single possible typo...

But in any case, Jean, if you want to make your database public, if it's formatted as a spreadsheet, then the best way by far would be to upload it as a Google document! 😄 It's easy to automatically convert a spreadsheet file into an editable (or just viewable, if you'd prefer) Google spreadsheet. That way everyone will be able to see it (or, if you prefer, only the people you select, although IMO the more open the better) and/or edit it. You can even give editing permissions only to certain trusted people/emails, if you fear vandalism. I'm happy to teach you how, if you need help!

The group could then do the metadata tracking work on the Google doc, which could be informed by the tag group(s), while keeping the discussion free and open here on Talk.

How does that sound? Am I making any sense? 😃

Thanks for being so clear and thorough about what you need! It might not be possible to keep all of that metadata in any version of Talk (except as a big list in a topic, which isn't too elegant), but there's no reason we can't use other popular and useful software as accoutrement, right?

P.S.: One more thought: Whereas Talk can be iffy at searching for hashtags outside of tag groups, the science team should have access to the actual database of tags, and might be able to pull what the Search currently can't.

Posted January 30, 2015 6:22 PM
by JeanTate in response to DZM's comment.

Thanks DZM.

but do you really expect over 500 candidates even at your first screening level? SDRAGNS are super rare, right?

Here's a conservative case that the number of candidate SDRAGNs in the FIRST/WISE part of RGZ (full database) is well over 500:
- SDRAGNs discovered to date, z < 0.1: 10
- of those, number in SDSS footprint: 5
- redshift out to which SDRAGNs will be POs (photometric objects) in SDSS: 0.5
- z=0.5 comoving volume/z=0.1 one*: ~100
- so, number of SDRAGNs in SDSS footprint, 0.1 < z < 0.5: ~500
- average number of candidates per 'the real thing': 10
- so, number of SDRAGN candidates in the FIRST/SDSS part of RGZ: ~5,000
Sure, that's a BOTE (back of the envelope) calculation; however, I think it'd be a serious challenge to get the estimated number of candidate SDRAGNs down to below 500.

Independent confirmation: in my database, my internal record number is up to ~350**. And less than half of the FIRST/SDSS RGZ database has been offered up to us zooites for classification (to date). Not to mention that I'm waaay behind in checking what others have posted/commented on/added to their Collections. Nor that only a handful of RGZ zooites even writes a comment/starts a Discussion/adds to a Collection ...

How does that sound? Am I making any sense? 😃

I'll try reading this again, later, when I'm not so tired; however, on first read it didn't make much sense (sorry). Quick question: can zooites who are in China use/access Google documents?

*thank you Ned Wright

**of course, perhaps as many as 90% of these will quickly turn out to be not excellent/good candidates, and over 50% pretty clear no/not candidates ... but you can't tell until you take a careful look, right?

Posted January 30, 2015 10:57 PM
by DZM Zooniverse Team in response to JeanTate's comment.

Hi Jean, let me see if I can explain a bit more clearly:

When you upload to a Google doc, you are basically putting your file online, but people can continue to edit it. You can decide who can see it and who can edit it... for instance, maybe you want everyone to see it, but only a few trusted people can edit it. Those people can all edit the same document that exists online... which is why it is so useful for collaborative work!

So the idea would be, use the Google doc as a dynamic database, but continue to talk about it here on Talk, where the zooites play. 😃

Again, I'm happy to teach you how to set this up!

I'm not 100% sure about China. I did some cursory research, and it looks like it's a bit of a toss-up whether or not a Google Doc will be viewable by a zooite behind the Great Firewall without using a VPN or other trick. Do we have a large Chinese-based contingent here on GZ:R?

Posted February 2, 2015 6:58 PM
by JeanTate in response to DZM's comment.

Thanks DZM.

From an Open Science perspective, putting files up, as Google docs, which anyone can read but only a trusted few can edit would seem to be the way to go; you don't want anyone to be able to edit, that sounds like an open invitation to spam*. I assume changes can be tracked, and rolled back if necessary (say someone inadvertently deletes half the data)? I'm not that familiar with Google docs; can records (in a spreadsheet/database) be URLs, say to JPEG images? Or JPEG files themselves?

Again, I'm happy to teach you how to set this up!

Thanks. I'd like to learn. However, I'm beginning to think that this isn't going to fly, either as a open, zooite collaboration leading to publication of a paper, or as a scientifically interesting topic (despite Ray Norris' words of "So keep your eyes peeled and yell out (very loudly) if you find one!", soon to be 14 months' old).

Do we have a large Chinese-based contingent here on GZ:R?

I have no idea. However, one of the most interesting SDRAGN candidates was discovered by Tony Wei, who at the time was in China (I've no idea where he is now).

*or worse, can Google docs be infected with malware?

Posted February 3, 2015 3:45 PM
by DZM Zooniverse Team

I've never heard of Google Docs getting malware. Google is probably the most tech-savvy company in the world. I wouldn't fear that. I would, however, nevertheless, make a backup copy for your own purposes every couple of weeks, or so!

I'm sorry to hear, though, that this particular ambition might not fly... ?

You're absolutely right about Google Docs being fantastic for open science. Let me know if you'd ever like my help in learning more about how they work!

Posted February 4, 2015 10:26 PM
by ivywong scientist, admin in response to DZM's comment.

Hi @DZM, RGZ is available in Chinese ...

Posted February 6, 2015 7:59 AM
by DZM Zooniverse Team in response to ivywong's comment.

True. And that's good! 😃 We want as much participation as possible!

But, unfortunately, the PRC's internet censorship (the "Great Firewall of China") can make it tricky to utilize some extremely useful tools for open science, such as Google Docs, while in mainland China. No matter how good we make Talk, it will always be better if it is supplemented by other software (for instance, we don't host images on Talk itself, so they need to be uploaded elsewhere before being displayed here).

So any project employing Google Docs might be off-limits to zooites in China. I was responding to @JeanTate's concern, trying to figure out if anyone here on Talk who might be a part of such a project would be affected, and if there's an alternative service that would work equally well?

Posted February 6, 2015 5:51 PM
by JeanTate in response to DZM's comment.

Here's a test: SDRAGNhunt01.

It should be a Google Sheet, with 34 rows (33 records); it should be open to anyone who clicks on it to read, but not to edit.

In it are several different kinds of records, for example:
- excellent candidates posted (and even with their own threads)
- good/fair/poor/no candidates, also posted
- not rated
- not posted
- knarly ones, e.g. outside SDSS footprint
One immediate issue: some 'when' fields were automatically converted to a mm/dd/yyyy hh:mm:ss format, even though the original record was not in this format* . Almost all these are copy/pasted from the date/timestamp in the relevant RGZ Talk Comment/Discussion post. So, how to tell Google Sheets to desist with the unnecessary (and possibly erroneous) conversion?

Next I'll try editing the document to add URLs and images; the former should be OK (they're just text strings), but the latter?

*while all the conversions are correct here, who's to say they will always be?

Posted February 9, 2015 4:05 PM
by JeanTate in response to DZM's comment.

About 2.5 years' ago, while in Beijing Karen Masters was interviewed by someone from 天文爱好者 (Amateur Astronomer, a magazine; main page here), and wrote about it in the GZ blog, A Bit More on the Chinese News about Galaxy Zoo. It would seem that there is considerable interest, in China, in doing astronomy of the online, crowd sourcing, citizen science kind. And it's cool that some of the various Zooniverse astronomy projects have Chinese versions, in either traditional or simplified form (or both?).

However, I wonder if the potential interest, in China, is far greater than the actual participation we've seen to date (not just in RGZ, of course)? Maybe someone (you, DZM?) could reach out to Karen, and ask her to facilitate contact with 天文爱好者?

Posted February 9, 2015 4:22 PM
by DZM Zooniverse Team in response to JeanTate's comment.

Hi Jean,

WOW! This looks fantastic! As you said, I can read it, but not edit it. You're killing it here! Thank you! 😃

Images--images are possible, but a bit unique. If you want them to appear in individual cells, it looks like you need to use a little trick. Here's a guide for a good way to do it. (Note, of course, that unless they are very tiny images, they will end up stretching the rows quite a bit!)

I think that the date conversion is related to the "formatting" of the cells. If that formatting bugs you, try highlighting that column (Column F) and then playing around with the options under the "format" tab (at the top of the screen). You should be able to select "plain text" to eliminate all formatting. You'll also see a number of options for different ways of displaying "dates." Let me know if this works? (You may have to turn it to plain text and then manually import the data for that column from your old spreadsheet.)

Again, thank you!

I can reach out to Karen if you like. Are you saying, just to facilitate more general interest in China and tap that potential market of citizen scientists? It seems like, at least in the short term, we can use this Google Doc until we get a report that a Chinese user wants to participate; then we'll work around it. Many tech-savvy Chinese folks also use VPNs to circumvent the government firewall.

Posted February 9, 2015 5:45 PM
by JeanTate in response to DZM's comment.

Thanks DZM.

I'll have a go at editing the Google Sheet, to fix the formatting, and include URLs. I don't see much point in adding images, though perhaps I could create a public (view only) album, using something from Google? That way I could put all the images in one album, and provide link(s) to it(them).

I can reach out to Karen if you like. Are you saying, just to facilitate more general interest in China and tap that potential market of citizen scientists? It seems like, at least in the short term, we can use this Google Doc until we get a report that a Chinese user wants to participate; then we'll work around it. Many tech-savvy Chinese folks also use VPNs to circumvent the government firewall.

Sorry, I (and this) was getting off-topic. I did mean facilitate general interest (there's already a connection, why not leverage it?), not find a way for our colleagues in China to get access to Google Sheets (etc); the latter is something I think is best left to them to consider and work on.

Posted February 9, 2015 9:00 PM
by DZM Zooniverse Team in response to JeanTate's comment.

I only talked about methods for adding images 'cause you asked. 😃

You can create a folder in Google Drive and upload all of the images there. Then you can change permissions for the entire folder, and, yes, provide links. Probably the best way to do it, if just linking to the object pages isn't enough.

Thanks again for your hard work on this!

Posted February 9, 2015 10:00 PM
by JeanTate in response to DZM's comment.

I've edited the Google Share document, converting the 'when' fields to plain text, and adding a URL*. Would you - any reader - mind checking that you can see these please?

You can create a folder in Google Drive and upload all of the images there. Then you can change permissions for the entire folder, and, yes, provide links. Probably the best way to do it, if just linking to the object pages isn't enough.

I did this, for the overlay images I have created and already posted, with the sharing permission for the whole folder being 'read only'; it's here ... can you see it?

This data dump is only 33 records, and 24 posted overlays; however, I think associating a candidate with an overlay image is already a challenge, and would be close to impossible if there were several thousand, say. But unless you can make the association, what's the point of having all the data+images up on Google Sheet+Drive?

I only talked about methods for adding images 'cause you asked. 😃

True, and until I really looked into it, I had nothing more than a vague idea of what might work. 😛

Here's a sorta idea of what I vaguely had in mind:

This is a screenshot, of part of a page generated by SDSS CrossID for DR7. The magic is in the SQL query, which I'm more than happy to share (big THANK YOU to zooite gumbosea for teaching me how to write such queries); as you can see, it's a sorta table, with cells containing a single text string ('name' and 'objID'), an image ('pic'), and a mix of text strings, URLs, and formatted data (the last column).

Random thought: suppose I had done all this, putting all my work on SDRAGN candidates up on Google Drive, and then the proverbial bus ran over me. Google, being Google, would immediately know (right?), and my Drive account would be frozen/deleted, right?

*since the unique URLs of posts in Talk Discussions cannot be found, these are simply the URLs of the ARG object; Talk will indicate which Discussion threads contain these fields, but - as we know too well - if the particular post is in a 20-page Discussion (say), as many/most of these are, good luck with finding it!

Posted February 10, 2015 1:31 PM
by DZM Zooniverse Team in response to JeanTate's comment.

Yes, I can see the changes in the Google Sheet (sheet, not share, FYI 😃 ) and the URLs work perfectly!

I did this, for the overlay images I have created and already posted,
with the sharing permission for the whole folder being 'read only';
it's here ... can you see it?

I can indeed, and it looks gorgeous. Look at all those candidates!

Random thought: suppose I had done all this, putting all my work on
SDRAGN candidates up on Google Drive, and then the proverbial bus ran
over me. Google, being Google, would immediately know (right?), and my
Drive account would be frozen/deleted, right?

No, Google doesn't (yet) monitor seven billion peoples' vital signs and attach that to their accounts. 😃 If something awful happened, your info would just sit there. It wouldn't be editable, though, but it would be possible to anyone make a copy of the spreadsheet and download/reupload all of the files so that science could continue. So I think your hard work is now safe!

Posted February 10, 2015 6:57 PM
by JeanTate in response to DZM's comment.

Thanks DZM.

If you check again today, you'll see a new column, "Image": it contains the name of the file (image) in the folder that 'goes with' the record (row). And while this does serve to make the association between image and record clear, it's clumsy, highly inefficient, and error-prone; quite unsuitable for more than ~a few hundred records.

No, Google doesn't (yet) monitor seven billion peoples' vital signs and attach that to their accounts. 😃

Perhaps. However, Google surely know more about me than even I do, given that everything is cross-linked within an account, gmail, use of Drive/Sheet/Docs, browsing history (if I use Chrome), ... not really 7x10⁹, rather the, what?, ~10⁸ Google account holders. 😛

Posted February 11, 2015 9:51 PM
by DZM Zooniverse Team in response to JeanTate's comment.

It looks to me like you've well gotten the hang of that sheet, now. Good work!

What you do with it is up to you, Ivy, the science team, and your fellow GZ:R zooites. 😃

Let me know if I can help you out with anything else!!

Posted February 12, 2015 4:23 PM
by JeanTate in response to DZM's comment.

Thanks DZM.

When there's a Zooniverse-wide discussion forum, perhaps we could have a broader discussion on what, realistically, zooites who get enthusiastic about a project can do, beyond just classifying?

Posted February 22, 2015 12:28 AM
by DZM Zooniverse Team in response to JeanTate's comment.

I think that's an excellent idea. I know that we have some incredibly smart citizen scientists here who want, yes, to classify plenty, but also to go beyond... and who have the know-how to do so!

It may differ from project to project, and citizen scientists may always be slightly hampered by not having direct access to top-notch tools for follow-up research. I'm not sure, myself.

However, there are also many projects from which we may be releasing completed data sets soon... there may be work that can be done with those!

The scientists will often know best, so in addition to the Zooniverse-level discussion, it will be good to have this discussion on a project-by-project basis, too.

A final thought... we heard a presentation by Authorea the other day ( https://www.authorea.com/ ), which might be an extremely useful tool if Zooites ever did decide to pursue a collaborative paper of some sort. Definitely an alternative to using Google Docs. They also may soon have some tools for helping first-time paper writers!

Posted February 23, 2015 7:16 PM
by zutopian

I revisit this old discussion in order to inform about a new paper.:

A Search for double-lobed radio emission from Galactic Stars and Spiral Galaxies
Abiel Felipe Ortiz Martínez, Heinz Andernach

We present a systematic search for two types of very unusual astronomical objects:
Galactic stars and spiral galaxies with double radio lobes, i.e. radio emission on opposite sides of the optical object, suggesting the ejection of jets from them. We designed an algorithm to search for pairs of radio sources straddling objects from two unprecedented samples of 878,031 Galactic stars from the Sloan Digital Sky Survey and 675,874 spiral galaxy candidates drawn from the recent literature. We found three new examples of double-lobed radio stars,
while for the spiral galaxies we only rediscovered one known such double source, confirming that the latter objects are extremely rare.

(Submitted on 8 Oct 2016)
https://arxiv.org/abs/1610.02572

PS: H. Andernach is co-author!

Posted October 12, 2016 8:45 PM
by ivywong scientist, admin

Thanks @zutopian. Indeed, confirmed double radio lobes from spirals are very rare.

Posted October 14, 2016 10:24 AM