example reproducible pipeline with a NetLogo wolf-sheep predation model and R Markdown
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
data
images
netlogo
packrat
results
rstudio
src
.Rprofile
.dockerignore
.gitignore
LICENSE
Makefile
README.md
docker-compose.yml
wolf-sheep.Rproj

README.md

wolf-sheep

ASU ABM club adaptation of the NetLogo wolf sheep predation model with added kill tracking from @mmannin5

Preparation

Software Requirements: docker, make and bash (That means old versions of Windows (< 10) definitely will not work without cygwin or creating run.sh in a different language. Linux and Mac will work fine).

  1. Install docker
  2. Clone this repository
  3. Build the Docker images by running ./run.sh build on the command line in the root of the project. Alternatively, images can be downloaded from DockerHub (docker pull comses/resbaz-analysis:3.3.3 and docker pull comses/resbaz-netlogo:5.3.1)

Run the workflow

Executing% ./run.sh run on the command line (caveat: only tested on Linux and Mac, it may run on Windows 10 bash) will execute the entire pipeline which currently consists of (1) a NetLogo behavior space experiment that generates a single CSV file, and (2) a RMarkdown file that loads the generated CSV file and generates a RMarkdown HTML report.

The NetLogo data can be found at data/vary_food_gains.csv and the resulting RMarkdown report will be produced at results/wolf_sheep_AB.html, relative to the root directory.

Interact with the workflow

  1. Run make interact
  2. NetLogo GUI should have popped up on your desktop
  • If it hasn't check the logs (docker-compose logs -f netlogo). If there is an error about not being able to connect to the X11 server window you need to give docker permission to talk to the X11 socket. This can be done with xhost +local:docker.
  1. RStudio GUI available at localhost:8787

Interact with NetLogo

The root project directory is mounted at /code so the NetLogo model used to generate this model is at /code/src/wolf-sheep-predation.nlogo. Open /code/src/wolf-sheep-predation.nlogo in NetLogo to view the model. The Behaviour Space settings used the vary_food_gains experiment. Parameters for the experiment can be seen by going into Tools > BehaviourSpace and opening vary_food_gains

Interact with RStudio

At localhost:8787 login as rstudio with a password of rstudio. The root project directory (the directory this is in) is mounted into RStudio at /home/rstudio/code. It should be visible in the explorer pane in RStudio.

  1. Open the wolf-sheep.Rproj project file in RStudio (File > Open Project... and select wolf-sheep.Rproj it will at the path /home/rstudio/code/wolf-sheep.Rproj). This will load the Packrat repository used for this project.

  2. Open src/wolf_sheep_AB.Rmd

  3. Run the knit command

    Location of Knit button on RStudio

Notes

This workflow has hard coded the output file generated from NetLogo into a path made accessible to the RMarkdown container. This example uses the hard coded data set name vary_food_gains.csv. A more robust workflow would define this type of data in a single place and referencing it in dependent computations so that changing the filename of a dataset does not require manually editing multiple files.

Packrat Resources

References

Wilensky, U. (1997). NetLogo Wolf Sheep Predation model. http://ccl.northwestern.edu/netlogo/models/WolfSheepPredation. Center for Connected Learning and Computer-Based Modeling, Northwestern University, Evanston, IL.