The below blog post was written by Alex Todd, an Ogden Summer Intern who spent the summer working on Galaxy Zoo related research projects at the University of Portsmouth. Alex is now off to his next adventure – starting his undergraduate degree in Natural Sciences at the University of Bath.
I have been working with the Galaxy Zoo team at the Institute of Cosmology and Gravitation, in Portsmouth, for 8 weeks this summer. I have been analysing the results of Galaxy Zoo 2, and more specifically the region of the sky known as Stripe 82. In this area, the Sloan Digital Sky Survey (SDSS) took many images of the same patch of sky, instead of only one. These images were combined to produce a single, higher quality image, which showed fainter details and objects. Both these deeper images and the standard depth images of stripe 82 were put into galaxy zoo, and I have been comparing the resulting classifications. I learned to code in python, a programming language, and used it to produce graphs from the data I downloaded from the Galaxy Zoo website. I started by comparing the results directly, comparing the number of people who said that the galaxy had features in each of the image depths.
On the graph, each blue dot is a galaxy (there are around 4,000) and the red dashed line shows the overall trend. As you can see from the graph, when the proportion of people who see features is low, there is a good match between the two image depths. However, when the proportion is high, there is a much bigger difference between the two image depths, with the proportion being higher in the deep image. This is because fainter features are visible in the deeper image.
I then plotted graphs of the difference between the proportions (P(Features)) against the brightness of the galaxy. To measure the brightness, I used the apparent magnitude, a measure of how bright the galaxy appears to us (as opposed to how bright it actually is).
The graph below shows the difference in P(Features) plotted against the apparent magnitude. The blue line is at y=0, and the green line represents the average value of the difference between P(Features). As you can see, there is not much difference between the values of P(Features) when the galaxy is particularly bright (Small apparent magnitude) or when it is particularly dim (large apparent magnitude). However, when the galaxy has an average brightness, the difference is quite substantial. We think this is because in bright galaxies, features can be seen in both images, whilst in dim galaxies they can be seen in neither. In medium brightness galaxies, however, they can only be seen in the deeper image. The fact that there are differences between the classifications means that it would be a good idea to classify deeper images of the rest of the sky, to hopefully improve the accuracy of the classifications.
I have greatly enjoyed my time working on at the ICG on galaxy zoo, and would certainly seize the opportunity to pursue it further.
It’s been a pleasure working with Alex this summer. He really impressed me with the speed at which he picked up programming languages. This information about the differences in perception of morphological features between deeper and shallower images is very useful to us as a science team as we plan for future generations of the Galaxy Zoo project with new, more sensitive images from current and ongoing astronomical surveys.