Help in using the Data for research project.

Tanay Singh
  • 6 Feb '23

Good evening,
I am Tanay Singh, a 3rd year Undergraduate Computer Science Student
studying at PES University, Bengaluru, Karnataka, India.

For our pre-final and final year research project, me and my teammates
decided to explore the applications of CS in Astronomy and Astrophysics. By
digging around the internet for some topics, we arrived at the following
problem statement for our research project : "Exploring the Galaxy-Halo
Connection through Galaxy Morphology: A Deep Learning Approach using
GalaxyZoo and IllustrisTNG".

We have figured out the first part of our topic, which is the Galaxy Morpholgy
bit using the GalaxyZoo Dataset and have reached a standstill point due to the
second bit of our topic, which is the Galaxy-Halo Connection.

With respect to this, I would like to place a few questions in front of you :
1) We will be using the TNG100-1 simulation data for our research. Is it
necessary for us to download the entire dataset for the same?

2) I was exploring the use of the API tool on your website using Python and it
seemed good for our topic, but my question is that if it is possible to access the
respective halo data only for the galaxies we will be using for our project? If yes
could you also provide the endpoint call for the same?

3) Is our approach this problem statement correct, or do we need to refine it a
bit more or change a few things here and there. I ask this question since we are
still lacking a bit on our domain knowledge part and while we are working on it,
a little suggestion from experts in this domain would do us wonders.

Thanking you for patiently reading through this entire query and I will be
eagerly waiting for your response to the same.

Thanking you once again,
Your Respectfully,
Tanay Singh

Dylan Nelson
  • 7 Feb '23

(1) You do not need to download the entire dataset. You can download only the catalog, only "cutouts" of the snapshots, and so on. You can also use the Lab service, to access the data directly.

(2) I'm not sure what you mean by "halo data" - do you mean properties of halos available in the catalog, or do you mean the "particle-level" i.e. snapshot data for a given halo. In the second case, then yes you can download this data just for a specific halo, i.e. i would look at [base]/subhalos/{id}/cutout.hdf5?{cutout_query} under the API documentation.

(3) The title combines many aspects. You can definitely explore the galaxy-halo connection with TNG, and you can also connect to morphology. However, I am not sure how you plan to connect GalaxyZoo to TNG - of course, the galaxies in GalaxyZoo are different from those in TNG. If you mean e.g. you will develop a ML technique to predict halo mass given galaxy morphology, using TNG as the training data, and then apply this tool to actual galaxies available in Galaxy Zoo, then yes this would be possible, but it is a big project.

  • Page 1 of 1