chumbley/fco repository overview

⁠A simple example analysis using Bayesian parameter ranking

• The docker image is hosted here⁠.
• A permanent, citable code archive here:
• The simple example, in the directory /fco/R does linear regression then asks for the most detailed but plausible ordering of parameters (by magnitude).
• Some useful bash scripts for shortcuts of docker commands are in directory /fco/bash.
• The submitted manuscript is here⁠. If you publish with this method, please cite the paper as:

Chumbley JR (2019) A Bayesian credible set for ranking parameters by relative magnitude. Manuscript submitted for publication. https://chumbleycode.github.io/docs/papers_reports/fcr_apa.pdf⁠

• For related work, see my website⁠. For example,

Chumbley, J.R., Potente, C., Xu, W., & Shanahan, M. (2019) A Bayesian approach to life course epidemiology. Manuscript submitted for publication. https://chumbleycode.github.io/docs/papers_reports/range_apa.pdf⁠

⁠1. Pull docker image

Run the following from the shell:

docker pull chumbleycode/fco:latest

⁠2a. To open container with rstudio in your browser

docker run --rm -p 2222:8787 -e USER=guest -e PASSWORD=secret chumbleycode/fco:latest

Then go to http://localhost:2222⁠ in your favorite browser.

You can change USER and PASSWORD, in the above. USER=guest PASSWORD=secret

⁠2b. Alternatively, to use R in the terminal

docker run --rm -it chumbleycode/fco:latest R

⁠3. Play with the example analysis script

A simple example script is in fco/R/example_analysis.R

⁠Alternatives to 2a/2b: Running container with local volume and root access

If you want to use the functionality on own data, you must additionally bind mount your local folder. First use the shell to navigate to your data folder then - for shell or browser access - run either

docker run --rm -p 2222:8787 -e PASSWORD=secret -v $(pwd):/home/rstudio/fco/host -e ROOT=TRUE chumbleycode/fco:latest

or

docker run -it -v $(pwd):/home/rstudio/fco/host chumbleycode/fco:latest R