dfm / exopop Goto Github PK

Inferring the population of exoplanets

License: MIT License

Python 21.52% TeX 78.48%

exopop's Introduction

Exoplanet population inference

This repository contains the code and text for the paper Exoplanet population inference and the abundance of Earth analogs from noisy, incomplete catalogs by Daniel Foreman-Mackey, David W. Hogg, and Timothy D. Morton and submitted to ApJ.

The code lives in the code directory and the LaTeX source code for the paper is in document.

Code

The meat of the probabilistic model is implemented in population.py. Then there are a set of scripts that you can use to generate the figures from the paper. You should look at the docstrings for details but the summary is:

simulate.py generates synthetic catalogs from known occurrence rate density functions,
main.py does the MCMC analysis on either real or simulated catalogs, and
results.py analyzes the results of the MCMC, makes some figures, and thins the chain to the published form.

Results

Our simulated catalogs and results are available online on figshare.

Attribution

This code is associated with and written specifically for our paper. If you make any use of it, please cite:

@article{exopop,
   author = {{Foreman-Mackey}, D. and {Hogg}, D.~W. and {Morton}, T.~D.},
    title = {Exoplanet population inference and the abundance of
             Earth analogs from noisy, incomplete catalogs},
  journal = {ArXiv --- submitted to ApJ},
     year = 2014,
   eprint = {1406.3020}
}

License

The code in this repository is made available under the terms of the MIT License. For details, see the LICENSE file.

exopop's People

Contributors

Stargazers

Watchers

Forkers

imclab ruthangus joshspeagle psarkis

exopop's Issues

Reading through latest version now

Will send another marked-up PDF asap.

make a proper README at the top level of this repo

audit for capitalization of Earth, Solar, and Sun.

I have rules

title, abstract, intro

I just refactored the abstract and it needs checking. I edited in the browser, so I may have broken LaTeX. I made huge changes to the ordering, to get it into context, aims, method, results order. I explicitly called out the differences with Petigura, which I now think is necessary.

I only tweaked the Intro. Let me know when you need more. Signed, @davidwhogg

write a few-sentence description / criticism of http://arxiv.org/abs/1406.6048

this might get incorporated to the paper at respond-to-referee

compare posterior inferences about small planets

As per my reaction to Scott:

We should compare the posterior inferences about the small planets in Petigura to those that emerge (implicitly) in our hierarchical model. These might be significant for the most important planets.

finish discussion and give to Hogg for once over

post on arXiv

Buy @dfm a new laptop

return to multiplicity issues in the discussion

I like the intro discussion of this, but at least mention it at the end too. The discussion should contain all the reasons we might be wrong, and, where possible, the direction that wrongness would take us.

modify paper and write response to referee

submit to AJ

or another journal?

fig 10 missing

no comparisons.pdf is in the repo; please add; I get a failure on make
on that figure (and in the text) replace "negligible uncertainties" to "uncertainties ignored" or "uncertainties not used".

get co-author sign-off and re-submit revised verison

write note about optimal estimators

If you really have a result about 1/Vmax weighting, this should be published as a separate note, either arxiv-only or in an appropriate journal

thank Eric Ford

either in the list with Bovy, or else explicitly as the key organizer of SAMSI

Stellar parameters

@benmontet thinks that we should discuss the effects of Bastien et al. (2014)'s result that the stellar radii are, on average, underestimated.

plot adjustments

What about putting error bars on the points? too busy?

Should day T [days](as it does) but then ln(T / day).

Also, the b/w printer will suck -- you have vibrating red on grey. Make the Petigura points full black (they are the data!), and stretch the greyscale from 0.5 to 1.0, not 0. to 1.0. Then you can make the completeness contours thinner too.

resubmit

I sign off

write abstract

Once you have a draft abstract, give it to @davidwhogg for a once-over.