This is a template for creating a Jupyter Notebook in R, originally created for NOAA RESTORE's:
Oysters, Blue Crabs, and Spotted Seatrout: Building Resilience to Environmental Trends and Variability in the Gulf of Mexico
Using GitHub and Binder, share your work:
- Describe your project
- Add Lat/Lon information to make an interactive map of the study area and for station data
- Do plots and analysis
In this Notebook, we create an interactive map of station data, plot a time-series of temperature and salinity for a sample station, then download data and make plots for additional stations.
Click the Binder button to open the sample Notebook:
This example repo uses R. (Using Python? Check Here.)
NOTE: Binder takes a very long time to spin up for the first time when using R. It should be faster the next time, if the image is still cached.
Starting from the landing page of this repo, click the green button Use this template then choose Create a new respository.
Choose the Owner
- If your project is NOAA RESTORE, please change Owner to OyBcSt.
- If you don't see OyBcSt as an option, contact Lisa to be added as a collaborator.
Choose the Repository name
- Choose something that indicates the type of project, e.g., hydro, climate, fisheries, etc.
README, scripts, and plain text files can be edited directly from GitHub by clicking the pencil icon, then click the green Commit changes button at the bottom of the page.
Change the README:
- Go to README.md. Click the pencil, and edit it directly to describe your project.
- Change the address for the Binder button...remember, the button is pointing at the template. See below.
- For Reference: GitHub markdown syntax
How to get the Binder button code:
- Go to https://mybinder.org (in a new tab).
- Paste the URL to your new repo.
- Put index.ipynb in the box for File.
- Where it says Expand to see..., click and copy the Binder button code.
Test the Binder button:
- Click the modified Binder button, and verify a working Notebook is being launched from your new repo.
Edits to the Notebook while in Binder are not saved to GitHub. When modifying the Notebook in Binder, periodically Save Notebook, then Download the file locally. Then the local files must be added back to GitHub.
To add local files back to the repo, go to the GitHub page, click Add file, then Upload files.
If you do lots of changes, uploading the files will get old fast.
Personally, I clone the repo locally, from Mac terminal:
git clone [https://address/reponame]
cd reponame
Then, I modify things while in Binder, Save Notebook, Download, and I download to reponame. Then, from Mac terminal, I do
git status
Check the status. The index.ipynb is probably the only thing that changed. And if it should have changed but didn't, you might want to save and download again. Check the download path. Then:
- git add .
- git status
- git commit -m "added more notebook"
- git push
If you were modifying the README.md online, you will need to pull those in before pushing out your changes:
- git pull
Contact me if you need help with git. We can develop some training and documentation based on what folks are having problems with.
If your Notebook is published and established, I think it is safer to modify things within Binder so you don't accidentally break your environment.
If you are developing a new Notebook, or need to make substantial changes, then I suggest...
For beginners:
- Download Anaconda
- Install the R kernal for Jupyter, see: How to use the R programming language in Jupyter Notebook
- Open and develop the Notebook in Jupyter with R as kernal (see above link)
- Find index.ipynb, or create a new file, modify, and save.
When the Notebook is ready, use
gitas above to push changes. Note: you can have several Notebooks in one GitHub repo.
Binder has to build your environment. If new libraries are needed for script modifications, you may need to change the runtime, or install new packages. That is what these files do:
- runtime.txt
- install.R
For Not Beginners:
- It is better to install miniconda, create a yaml file with dependencies, then
conda env create, for each project. - See this for more about Conda
- Unfortunately, I did not test running a Notebook with R using miniconda...but we'll figure it out.
Contact me if you need help with the environment. We can develop some extra documentation based on what folks are having problems with.
This work was funded by National Oceanic and Atmospheric Administration's RESTORE Science Program under award NA19NOS4510194. The modeling data is being hosted on Open Storage Network (OSN), through allocation EES210015 from the Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support, which is supported by National Science Foundation grants #2138259, #2138286, #2138307, #2137603, and #2138296.