Galaxy Zoo Starburst Talk

New Tools Tutorial

  • trouille by trouille scientist, moderator, admin

    Dear all,

    I know we’re all raring to go on the data analysis phase of this project. Tools.zooniverse.org will be invaluable in this process.

    We’ve created a special Quench platform within Tools. You select it by choosing ‘Quench’ within the dropdown menu next to ‘zootools’ in the upper left of the page in tools.zooniverse.org.

    To help you navigate this platform, we've created a step-through tutorial. When you first enter Quench Tools, the tutorial prompt will pop-up. Through the tutorial, you’ll learn how to manipulate the post-quenched galaxy sample data table and then create a scatter plot using the new column of data you’ve derived.

    Today we’ll be hosting a GZ google hangout to go through the tutorial and answer questions. We’ll start at 15:30 CDT (Chicago), which is 13:30 PDT, 16:30 EDT, 20:30 GMT, 21:30 BST, 22:30 CET, and 23:30 CAT. We’ll post the recording at http://blog.galaxyzoo.org/2013/07/23/next-gz-tools-hangout-wednesday-24th-july-2030-gmt/.

    Now go play! There’s so much functionality within Tools for you to explore. If you have a question or come across a cool result, please post!

    Also, please see the post about ‘Sharing Tools Dashboards’ for how to use Tools to facilitate the collaboration process.

    Posted

  • karthikeyan.d by karthikeyan.d

    I see that there is a guide for the Prompt commands. Can you add a guide for the Axis variables too? I followed the step-through tutorial and created a scatterplot with log mass in the X axis. However, I don't know what "log mass" is. (Mass in a log scale?) I understand some of the options but it will be useful to have a guide that explains the different options.

    Posted

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

    This is wonderful news!

    For myself, I'll be waiting until after the Hangout before I start playing ...

    Posted

  • JeanTate by JeanTate

    I'm interested to know what others thought about the Hangout yesterday (Wed 24th July).

    For me it was hard to follow - the screen shots were simply too fuzzy to see anything, I couldn't even see which item on the menu bar was being clicked most of the time! and the audio track was poor - so I got almost nothing out of it.

    Posted

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

    Full screen, with a little creative guess work, I could just about make out menus and buttons. However, it's all sufficiently intricate getting started that I'm sure I'm have to work through it with documentation. I felt it delivered half of what I was hoping to understand.

    I felt I came away with a good overview of the structure and capabilities of the data analysis tools. Where I'm coming unstuck is having a clear sense of what we're trying to demonstrate using the dataset. I guess it's conceivable that some may demonstrate relevant correlations but, without having researched the topic in more detail, I'm challenged by the prospect of linking this analysis to some of the stated questions such as 'what caused the quenching in these galaxies' or ' the role that post-quenched galaxies play in galaxy evolution'.

    As an aside is was interesting to put some faces to names to get a sense of who is working away behind the scenes.

    Posted

  • mzevin1 by mzevin1 scientist, moderator

    Sorry to hear it was hard to follow some of the Hangout. The biggest thing that we were trying to get across was that Zooniverse Tools will offer us a way to jump right in and analyze the data that we've been gathering by classifying the post-quenched and control sample without needing to be experienced in any computer languages. All the information from our galaxies is easily accessible through Tools and allows us to churn our diverse plots like crazy to see what works and what doesn't work. Once the classification process is complete we'll have morphological data to work with in Tools as well.

    The best thing to be doing with Tools right now is to go in and play! Tools is still in Beta, and many on the science team are new to Tools as well. There is now a tutorial as well on Tools that should give you guys a good jumping-off point in how it works. Before the classification is complete is a great time to explore what capabilities tools.zooniverse.org has for our disposal. Any questions about Tools can also be posted on this forum.

    James_involute brings up a good point - What are we trying to demonstrate with our data? When the classification process comes to an end, we're going make a good deal of background information (including summaries of a few important scientific articles on post-quenched galaxies) available to you. This whole project is letting everyone involved be immersed in the entire scientific process, and a big part of that process looking at background information on the topic. What do we already know? What new things can we learn? How can we expand on ideas that have already established? How can we do things differently?

    Keep up the good work classifying everyone! Exciting things to come.

    Posted

  • JeanTate by JeanTate

    OK, I'm running through the tutorial, and when I get to entering something in the Prompt field (and Executing), the message "Notice how the new column has been added to the end of the Table." appears ... but no new column has been added!

    I suspect there's something missing - or similar - in the "To create a column u_mag - g_mag color, type "New Field color, .u - .g" in the Prompt, and click "Execute"."

    Posted

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

    OK, it worked, when I did it again ... for the 5th (?) time!

    Posted

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

    How do you delete a field? Say you wanted to create ug (u-g) but accidentally created ug (g-r); how do you delete the field ug?

    Posted

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

    I tried to create a New Filter; nothing seemed to happen! I assume that the number of records in my table should decrease (my new filter should have left me only ~10%). Help!

    Posted

  • james_involute by james_involute

    Recommendation - dispense with IE10 - I couldn't make any headway - I seem to be getting a great deal further using Chrome.

    Posted

  • vrooje by vrooje moderator, scientist

    I've encountered some bugs myself, so please do report them to the developers, and I agree that IE is not that reliable.

    By the way, I've now annotated the video, so hopefully some things are a bit clearer. I tried to add highlighting boxes to show when something important was being clicked, and I've added comments here and there as well. Let me know if something else needs adding... and I'm sorry it took this long, but it was a slightly slow editing process.

    Cheers,

    -Brooke

    Posted

  • trouille by trouille scientist, moderator, admin in response to JeanTate's comment.

    There is no way to delete a field. It's OK to make mistakes with those talbes. You can always start a new table, or just not use the mistake column in your next steps.

    Posted

  • trouille by trouille scientist, moderator, admin

    Thanks for annotating the video vrooje! Seems like that addition could be quite helpful.

    Posted

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

    Yes, it makes the video much easier to follow, as a way of introducing Tools. Thanks Brooke/vrooje.

    Posted

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

    Finally I have some time to play with Tools. 😃 I have the same issue as you Jean. I've tried 10 times now to create the extra column in the tutorial table with no luck - and before anyone asks I copied and pasted the command "field 'u - g color', .u - .g" to ensure the syntax was correct!

    Anyone else having this problem?

    Posted

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

    Now the fun really begins, jules! 😄

    After the first time, I did not try using the Tutorial again; rather, I created a new Dashboard, gave it a new name, loaded up my own data - either from one of my own Collections (you have to choose 'Zooniverse', even if your collection is in Quench!), or one of the two Quench sets (Sample or Control) - and went from there. So far today I've had no trouble ... except for what I've already noted (mostly to do with 'log' not working, and why the data seems to contain some, um, strange fields/values).

    Maybe it's a 'newbie bug', associated with the Tutorial, and only the first time a zooite tries it?

    Posted

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

    Sorry I'm just going through these now. You can delete filters and fields by clicking the 'X' next to where the statement appears underneath the prompt.

    Posted

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

    Thanks Jean. 😄 My data analysis skills are rather rusty but I'm happy to mess about with plots and post anything that looks useful. I think I'll bypass the tutorial too and just dive in.

    What can possibly go wrong...? 😉

    Posted

  • jules by jules moderator in response to edwardothegreat's comment.

    Thanks Edward! Though if the tutorial has the same glitch for all users I'm surprised at the lack of feedback. Unless, of course, people have taken the accepted approach of ditching the manual! 😮

    Posted

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

    Hey jules are including the double quotes when you type it in (you shouldn't)? It works fine in both chrome and firefox for me.

    Posted

  • edwardothegreat by edwardothegreat

    Ah I see what was meant about fields not getting deleted. I think I should have a fix for that early next week.

    Posted

  • jules by jules moderator in response to edwardothegreat's comment.

    Ah yes I was. Thanks for that - I'm not familiar with the syntax so I took the instructions literally! However, having removed the double quotes it didn't work straight away - only on the fourth attempt! Copy and paste again so the command was the same all 4 times.

    Ah well - onwards and upwards...

    Posted

  • edwardothegreat by edwardothegreat

    It sometimes takes a while to apply the prompt command, since I definitely did not anticipate having as much data as quench does in the dashboard at once. Later next week I should have a way to apply it more quickly.

    Posted

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

    I noticed this too - the delay. Maybe it would help if the 'training set' were a bit smaller? Say ~100 objects?

    Posted

  • Peter_Dzwig by Peter_Dzwig

    OK. Can I reload the tutorial?

    Posted

  • trouille by trouille scientist, moderator, admin

    You'll know a 'prompt' within the Quench Tools table has worked if below the 'Execute' and 'Guide' buttons, a new line (with similar words as the prompt you just submitted) shows up on the screen. Also, as you mention, if you've applied a filter, the table should have fewer rows.

    Try the following prompt:
    filter .u > 15

    Does that work?

    Posted

  • trouille by trouille scientist, moderator, admin

    You can reload the Quench Tools tutorial by creating a new dashboard and labeling it Tutorial. Importantly, it only works if you've given it the specific name Tutorial. Calling it tutorial doesn't work, nor does TUTORIAL.

    Posted