Public data, code, and notebooks for S4G-based bar-frequencies paper using S4G data

This git repository contains data files, Python code, and Python and R Jupyter notebooks which can be used to reproduce figures and analyses from the paper "The Dependence of Bar Frequency on Galaxy Mass, Colour, and Gas Content -- and Angular Resolution -- in the Local Universe" (Erwin 2018, Monthly Notices of the Royal Astronomical Society, 474: 5372; arXiv:1711.04867).

The data/ subdirectory contains text-file tables with various data compilations and simulation outputs; see the README.md file there for details.

(This figure, reproduced from the paper, shows the fraction of spiral galaxies which have bars, as a function of stellar mass, for the local, S4G-based sample studied in the paper (red circles), as well as for several SDSS-based studies. The blue pentagons show what would happen if the same S4G-based sample were to be observed at redshifts typical of the SDSS-based studies, assuming that only bars with projected semi-major axes more than twice size of the typical PSF FWHM can be detected.)

Dependencies

The Python code and notebooks require the following Python modules and packages:

• Numpy, Scipy, matplotlib
• Astropy
• Michele Cappellari's LOESS code: cap_loess_1d

The R notebooks require the survey and zoo packages.

Jupyter Notebooks

There are three Python notebooks:

• s4gbars_main.ipynb -- generates the largest set of figures in the paper; also generates data files for use in R logistic regression

  • Figures 1, 2, 4, 5, A1, A2, B1, B2 [currently incomplete for some of the latter figures]

• s4gbars_barsizes.ipynb -- generates figures which use S4G (and sometimes Galaxy Zoo 2) bar sizes

  • Figures 3, 6, 7, 8, 9, and 11 of the paper

• s4gbars_simulated_surveys.ipynb -- generates figures using the output of survey simulations (which are themselves generated by the Python script make_simulated_surveys.py)

  • Figures 10 and 12 of the paper

There are also two R notebooks:

• s4gbars_R_logistic_regression.ipynb -- runs the logistic regression analyses used for the paper (e.g., Sections 3.1, 3.2, 3.3, and 6.1 and Table 3).

• s4gbars_R_quantile-loess.ipynb -- generates text files containing quantile LOESS curves for bar sizes, used in Figure 8 of the paper.

Python Code

• datautils.py, plotutils.py, s4gutils.py -- miscellaneous utility functions (including statistics).

• simulate_surveys.py -- code for generating bootstrapped mock surveys measuring bar frequencies, using the S4G galaxies as a parent sample and adopting user-specified redshift ranges.

• make_simulated_surveys.py -- executable script generating specific mock surveys using the code in simulate_surveys.py:

  • SDSS-style bar fractions as function of stellar mass
  • SDSS-style bar fractions as function of gas mass ratio
  • HST-style bar fractions as function of stellar mass

  The outputs of this script (using the default random seed value of 100) can be found in the data/ subdirectory.

• generate_GZ2-bar-sizes_table.py -- code for regenerating the GZ2 bar-sizes table in the data/ subdirectory (note that running this will required downloading the GZ2 SDSS metadata table from the GZ data site; see notes in the data/external/ subdirectory).

How to Generate Figures and Analyses from the Paper

1. Download this repository (some individual notebooks can be run with only a subset of the data files and code, but it's simpler just to work with the entire set of files).

2. Run the Python notebooks (s4gbars_main.ipynb, s4gbars_barsizes.ipynb, s4gbars_simulated_surveys.ipynb) to re-generate the figures, or to experiment with alternative versions.

3. To re-run the SDSS survey simulations, use the make_simulated_surveys.py script, which will regenerate the output sim_* files in data/ subdirectory (using the same random seed as was used for the paper). To change the random seed for the simulations, edit the make_simulated_surveys.py script and change the value assigned to the randomSeed variable; to use the current time as the seed, set randomSeed = None.

4. To re-do the logistic regression analyses, run the R notebook s4gbars_R_logistic_regression.ipynb.

Licensing

Code in this repository is released under the BSD 3-clause license.