2nd Blog Post: Visual Classification of Galaxies

This week, I have finished classifying the 696 galaxies and have classified 124 extra galaxies- which turned out to all be galaxies that were part of the first set. The idea of doing this was to classify some galaxies twice (without knowing that was what I was doing) so that I can compare the results from both classifications and see how they are different. I have found that it is quite possible to disagree with yourself. I have created some graphs to analyse the results from the classification. I have also performed a similar analysis on some results from the Galaxy Zoo project (for more information go to https://www.galaxyzoo.org), which contains classifications for about 900,000 galaxies, to give a far more detailed picture of the distribution of galaxies in the local universe.


This is a 3D graph showing the spacial locations of merging galaxies in the most densely populated (or sampled) region of galaxies. The magenta dots show galaxies that are not merging and the blue dots show galaxies that are merging. Earlier results had shown that this area of the superstructure contained the highest density of mergers, and at first glance it seems that within this volume most mergers are occurring in the area that appears towards the top of this image, but further analysis will need to be performed.


This graph shows the types of galaxies that were classified in the first sample (of 696 galaxies) in red, and the second sample (of 124 galaxies), in blue. 0 represents point-like galaxies, 1 represents elliptical galaxies, 2 represents disky galaxies, 3 represents irregular galaxies, 9 represents galaxies that are too faint to classify and -9 represents images that do not contain a galaxy. As you can see, more elliptical galaxies were classified in the sample of 124 galaxies than were classified in the entire 696 galaxies the first time. I expected this because I had changed the definition I was using to classify ellipticals, but the effect of this was more dramatic than anticipated.


Finally, this graph, made using the data from Galaxy Zoo, shows the effect of a debiasing process applied by the Galaxy Zoo team on the classifications made of ellipticals. The red bars show the biased results (those collected from people classifying galaxies) and the blue bars show the unbiased results (voted on by the Galaxy Zoo team), and the curves in the corresponding colours show the distribution of each set of results. This graph demonstrates that there is a tendency in inexperienced humans to not classify elliptical galaxies as such- the same tendency I showed in my first set of classifications.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s