Creating a Stan Ecology Page

Dear Stan ecology community,

In order to foster the community of Stan users in all areas of ecology (incl. quant/stat ecology), we will formally begin the process of creating a central online presence. To reinforce that this is a community effort, we would first like to hear from any of you who are already implementing Stan in their work! Please contact us at stanecology@gmail.com to:

  • Let us know about your work and how you’ve used Stan
  • Would you be willing to provide a case study demonstrating a complete Bayesian workflow in a particular area of ecological research? (We’ll work on a general template to describe what this would entail in the near future.)
  • Do you have any publicly available Stan-related resources?
  • Do you have any publications/general work/posters that have used Stan?

The Stan Ecology page will initially be created using the information obtained from the community members who respond, those who mention they would like to actively participate in this venture, and those who have already responded to a previous post on the Stan discourse page: Fostering Stan user communities through domain-specific resource pages Topics like Bayesian workflows, prior/posterior predictive checks, and incorporating domain expertise will be emphasized.

Anyone is welcome to send an email detailing what they would like to see on the Stan Ecology community page as well. For those who have already reached out, expect an email in mid to late January with specific details on the next steps.

Excited to hear from you all!

Vianey Leos(-)Barajas, North Carolina State University
Jonah Gabry, Stan Developer (@jonah)

15 Likes

Hi Vianey and Jonah

I posted a new paper to this forum on 17 Dec 2019 in which we use Stan to understand how cell-to-cell spread of virus (HIV) becomes sustained and established. I’ve copied that post here, and will send to the gmail address. Thanks!

Hello Stan Community,

How does a virus go viral? We recently published a paper to quantitatively define the population dynamics by which cell-to-cell spread of HIV becomes exponential, starting from one infected cell. This process is fundamental to enumeration of HIV-infected cells in HIV-infected individuals on effective treatment, as well as the processes of HIV transmission and viral recrudescence if treatment is stopped. As part of this paper, we used Stan to condition a statistical model (of viral population establishment) on experimental data generated at the NIH Vaccine Research Center. The paper is highlighted in the December 2019 issue of Cell Host & Microbe (https://www.cell.com/cell-host-microbe/current) with journal cover art generated with ggplot2 in R and a “Preview” written by another group in the field.

There is free access to this article until Jan 30, 2020 using this link:
https://authors.elsevier.com/a/1aD52_hwRXT5aY

Thank you to the Stan developers and wider Stan community for making Stan such a great platform to perform Bayesian inference.

Best regards,
Jason Hataye, MD, PhD
Division of AIDS
National Institute of Allergy and Infectious Diseases (USA)
jason.hataye@nih.gov

Citation from PubMed:
Hataye JM, Casazza JP, Best K, et al. Principles Governing Establishment versus Collapse of HIV-1 Cellular Spread. Cell Host Microbe, 2019;26(6):748–763.e20. doi:10.1016/j.chom.2019.10.006

2 Likes

Hi Vianey and Jonah

This is perfect!
I will send an email answering yours points.

Thanks!

1 Like

I use brms and Stan and lot in my work. I have some in-house tutorials planned I am interested in hosting. Also, I have a methods paper in the works that could be used to guide some developments in the software, since at the moment I think brms and Stan are the best packages available for implementing the methods.

1 Like

So excited about this! Hopefully the community grows quickly once there are more Stan tutorials for common ecology models. Cole Monnahan et al. had a really nice paper in MEE introducing HMC to ecologists and fisheries biologists. I think having tutorials of the full process including model diagnostics and summaries will really help bridge the gap for many people.

3 Likes

Thanks for taking this on. The only thing we have like this now is a page for Stan uses in education. Here are some references:

5 Likes

I love the Lotka-Volterra case study! I really want to do a D&D dungeon ecology model for Stan sometime @Bob_Carpenter .

3 Likes

Hey all, will be very interested in the central page. My Stan chops are growing but not really strong enough to contribute much. I could at some point (ie when I finish my current paper) contribute maybe a lower level style case study that would be accessible to people who are just getting started

5 Likes

We’re also into it; possible contributions could include testing and development of workflows, with a particular focus on microbial ecology.

3 Likes

That should be easy—there’s an alpha predator in every room!

2 Likes

This would be fantastic! There’s really no need to have a certain level of Stan expertise. It would be wonderful to have resources for Stan ecology users at all levels. I’ll make sure you’re included on the mailing list and conversations.

Hello!

I am an ecological stan user too! And would be interested to participate to such a page :)

I already gave some workshop using stan (Data generating process and Introduction to Hierarchial Models), and have two case studies available (one on the simpler case of JSDM and one on Dirichlet regression).

I would be happy to contribute and share my tests more widely.

Lucas

1 Like

I have used Stan and rstanarm quite a bit in recent years. I have a series of blog posts on selecting covariates in multiple regression paired with R notebooks at http://darwin.eeb.uconn.edu/uncommon-ground/variable-selection-in-multiple-regression/

Kent

1 Like