COVID-19 Spatial Accessibility Reproduction
Nov-02
: In this lesson, we will reproduce COVID-19 spatial accessibility.
The goal of this lab is to get accustomed to working in a cyberinfrastructure environment with Python, Jupyter notebooks, and GitHub integration, and understand how to implement spatial accessibility analysis in Python.
We will do so with the Jupyter notebook published to accompany Kang et al (2020).
The notebook has been published on CyberGISX and maintained in a GitHub repository.
We have published the results of our prior reproduction and reanalysis studies in this repository: https://github.com/HEGSRR/RPr-Kang-2020
By the end of lab, you should have:
- executed the complete notebook
- taken notes and added comments to the code cells and/or content to the markdown cells in the notebook on CyberGISX to understand how the code relates to the publication (i.e. which blocks of code implement which sections of the paper?)
- pushed your changed (annotated) version of the notebook back to your own GitHub repository
Instructions
- Let’s continue working in groups as we did for the Spielman et al replication.
Set up GitHub for Integration with CyberGISX
- Fork the Rpr-Kang-2020 repository
- Set up an authentication token for connecting GitHub and CyberGISX
- Go to your GitHub User account settings
- at the Bottom of the settings menus, find
Developer Settings
- go to Personal access tokens –>
Generate new token –> Generate new token (classic)
- You can give the token the name
cybergisx
- In terms of
scopes for this token, you only need the first scope: repo
- You will need this authentication token for communicating between CyberGISX and GitHub.
- Here is the GitHub documentation on personal access tokens
Set up CyberGISX
- Log into Google services (e.g. drive.google.com) using your Middlebury credentials
- Go to the CyberGISX Hub https://cybergisxhub.cigi.illinois.edu/
- At the top-right, of the page,
Register with CILogon and create and account / log in using Google and your @middlebury.edu Google account as your Identity Provider
- Fill out the registration form with
organization: Middlebury College, website link: your google site and class: Open GIScience
- Once you are logged in,
Launch CyberGISX at the top-right
Linking CyberGISX and GitHub
- Go to
Git and Add a remote repository to clone your Rpr-Kang-2020 repository into the CyberGISX servers using the web address of your repository on GitHub
- Try making a small change.
- E.g. open the main
readme.md, add your name to the list of authors, and save readme.md as a markdown file
- Open the
Git panel from the Git icon at the left
- The Git menu shows
changed files
- Click the plus button on changed files to
stage the change, moving the file to staged
- Write a
summary of changes at the bottom of the Git panel
Commit the changes with your GitHub name and email address.- At the top of the panel, an upload cloud icon will have an orange dot to
push the changes to GitHub.
- To
push, you will need your GitHub user name and authentication token.
- Check to see if the change(s) is/are reflected on GitHub.com
- Always remember to pull new changes into a work session as you begin and push important changes from a work session as you conclude!
Executing a notebook
- One advantage of using a JupyterHub server is that server administrators can set up multiple Python environments from which to choose. For this lab, let’s use the default
Python 3 environment.
- Switch to the file browser panel and find the
procedure\code\02_COVID-19_Acc-Original.ipynb notebook
- Verify that the
kernel processing environment for the notebook is Python 3 (at top-right corner of the notebook)
- Because this environment is already set up, you do not need to install any packages!
- Run the notebook. There are three decisions to make in order to create different outputs, so I suggest group members try different combinations of these decisions:
- Street network: buffered or not?
- Population: people over 50, or COVID patients?
- Service: ICU beds, or ventilators?
- Note how long the hospital catchment process took to run and enter results in this GOOGLE FORM
- Save results, and then open
procedure\code\03_COIVD-19_Acc-Reanalysis.ipynb
- Execute the code, trying differnet combinations of Population and Service
- Note how long the hospital catchment process took to run and enter results in this GOOGLE FORM
- Take notes on how this reanalysis different from the original notebook.
Transitioning to next week you should:
- Compare the code repository to the published paper and brainstorm ways in which the code repository could be improved with regards to computational efficiency/speed, reproducibility, uncertainty, errors, cartography, aesthetics, readability…
## Software References
Reference
Kang, J. Y., A. Michels, F. Lyu, Shaohua Wang, N. Agbodo, V. L. Freeman, and Shaowen Wang. 2020. Rapidly measuring spatial accessibility of COVID-19 healthcare resources: a case study of Illinois, USA. International Journal of Health Geographics 19 (1):1–17. DOI:10.1186/s12942-020-00229-x.
Main Page