Apr 18
Merger hunters needed
One of the most impressive aspects of Galaxy Zoo has been the way we’ve been able to use the data in all sorts of unexpected ways. Whether it’s finding overlapping galaxies or strange blue blobs it’s become obvious that there is a lot more to all of your efforts than just elliptical vs spiral. When we were putting together the site, we didn’t give much thought to the many galaxy mergers in the catalogue; as far as the primary science goals went they were mostly just contaminants in our data set.It quickly became obvious that we had the chance to assemble a large and interesting set of mergers and learn more about why, where and how these beautiful collisions occur; conducting the first investigation into this is Daniel Darg who’s based here in Oxford.
He’s already got interesting results, but we’d like his paper to be as comprehensive as possible and that means including as many mergers as we can find. Everything which has a weighted vote of more than 60% in the merger category has proved to be a true merger, but now we need for help to have a closer look at those which have a vote between 40% and 60% in this category. We’re keen to get on with this, so rather than wait for Zoo 2 we thought we could ask for your help here. To find out how, read on below.
Update : New, easier classification system available : See this post.
We’ll find a way to say thanks depending on how many people answer the call!Here’s what we need you to do.
1. Read Daniel’s guide to what makes a Galaxy Zoo merger, available as a pdf file.
2. Download a list of objects which have between 40 and 60% of weighted votes in the mergers category from the file list below. Each one contains 1000 objects. These lists will open in a text editor. They currently DO NOT work in Excel.
3. Use the SDSS explorer http://cas.sdss.org/astro/en/tools/explore/obj.asp?id=588295842857943202 (replacing 588295842857943202 with an id drawn from the list) to view the system, and decide whether it is a merger or not.
4. Make TWO files; Call the first YOURNAME_MERGERS and include the id of the galaxies you’ve checked and found to be mergers. Call the other YOURNAME_NOT and include the id of galaxies you’ve checked and have decided aren’t mergers. The files should either be a comma-separated list or an excel file. If you are producing a second set then just add 2 to the end of the filenames and so on.
5. Email us at galaxyzoomergers@gmail.com attaching both your files.
6. Sit back and relax.
Thanks a lot.
List of files -
I think there is a glitch Chris. The object id in the downloaded excel file is always represented as 5.88E+17 and is not recognised when one enters it in the SDSS id
Give me a few hours and I can do you a website that does most of the mechanical bits of this analysis for you.
Sounds like a brilliant idea Waveney; I’ll sit back and wait a while.
Graham - thanks for pointing that out; I’d forgotten the problems Excel has in handling large integers.
Waveney - yes please! Please email me if you need more details (or PM via the forum).
Let us know, I’m itchin’ to go do me some classifications!
Work just in progress.
To avoid the excel problem import the file in text cell.
I like mergers more than ellipticals, that’s for sure
Would you like the files as html with the links setup
already? It is easy to do?
As a 1000 entry http txt file, there is no way of eliminating the object IDs as the analysis progresses, in order to eliminate errors and back searching for the next object to analyze.
This is a terribly clunky way of doing something of this nature. I’ll pass, after having four windows open at the same time and hopping between them, then getting lost trying to determine the next object ID.
Good luck!
Is there any ‘dead-line’ for a first set of answer? (just that we don’t keep lots of data and miss when you need it as well as not mailing every 20 galaxies)…
altymczuk : Yes, it’s clunky, but for a one-off task with everyone concentrating on developing Zoo 2 it seemed like the best way to do it. If someone wants to generate a list of links, then please do.
Saint-Pierre : There’s no deadline, but remember it is useful if more than one person checks the same galaxies as that helps us understand the errors in the results.
A website for doing this is at:
My Server
Clunky is the word! I have opened two spreadsheets in Open Office, and formatted them as “Text”. This gives the object ID correctly.
I have copied the whole of the (first) list into a third spreadsheet, so I can copy & paste from that into the address bar, then to the Merger or Not sheet.
It is a time-consuming process, so I added the “preamble” of the ID address into a column, then Concatenated into a new column which then acts as a simple “click” onto the object itself. A bit fiddly, but I can go straight from that sheet to the object.
This will also work in Microsoft Excel itself.
Some of the ‘yes’ examples on the tutorial page seem to me as possible overlaps.
How to convert the text lists to hyperlinked lists with Excel:
1) Right click the text file link, select Save Target As tto somewhere you can easily find on your computer.
2) Open Microsoft Excel.
3) Open the text file from within Excel and complete import to create a single column A of text.
4) Click in column B next to the first entry, cell B2, and enter the following equation:
=HYPERLINK(CONCATENATE(”http://cas.sdss.org/astro/en/tools/explore/obj.asp?id=”,A2))
5) Press Enter.
6) Left click the lower right corner of B2 and drag down column until the formula has been copied into all of the cells next to the numbers.
7) You can now click on each hyperlink to go to the SDSS page.