Hunting Programs for all
I created the Merger Checking (which now has over a million clicks), three Pea Hunts (All finished) and now an unofficial irregular galaxy classification. What started as 80 lines of Perl code is now 800 (about 50 of the original 80 lines still survive), but can now support almost any Galaxy Zoo mini-project.These mini-projects will never be as pretty as the main GZ sites, but they are quick to build, modify and use.
The original program worked, but needed considerable tweaking to get the balance of image selection right (there are now three algorithms), the skyserver interface and results storage to work with lots of users. I added features to make the checking easier, and features I had always wanted in the original Zoo. The code grew and was bodged from feature to enhancement to bug fix.
To undertake the first Pea hunt, I replicated the merger checking code, tweaked it a lot and launched the hunt. The second Pea hunt was a very tweak to the first. But again it did require changing the code to handle the new task. I knew this could be significantly improved.
A month ago there was a discussion on the forum about a possible project for a student and a project on irregulars was discussed. To support this it was suggested that a simple filter should be run to remove the junk from the objects in the irregular topic on the forum. However, Jules suggested some really much better ways we could classify the irregulars, as we went through and removed the junk. I had prepared the quick filter using the copy and tweak method), but the student decided to do something else. A few weeks passed.
One day when bored, I thought about the classifications Jules had suggested and decided it was time to revise the hunting program to support that sort of task, also to make it data driven at the same time to make it very quick to configure and lastly to automate the switching between the image selection algorithms.
Now any hunting task just needs a list of objects, and the questions to be presented which is most of a very small configuration file, along with things like the project title. The rest is now automated.
I launched the irregular classification a couple of weeks ago just as a bit of fun, as people are waiting for Galaxy Zoo 2. The idea is that it is a community led project, as soon as there are enough results I will put them where others can read them and analyse them to see what they can find out. I don’t know what we will find out, but we are sure to find something. Originally this was over 2,500 images in the irregular topic, and then others where added from dwarfs, blues, oddballs, newbies and finally anywhere. Currently we have 4,000 images of possible irregulars being classified.
When the request came in for the third Pea Hunt I was ready. Setting this up using the new program was really trivial – while the server was pre-compiling the list of objects, I wrote the questions and prepared a configuration file, 5 minutes later the hunt started.
I started in the Zoo as a quiet clicker on galaxies. Today my count of galaxies classified is poor by the standards of many of the regulars. However, I have found myself a niche. Click away.