COVID-19 Spatial Accessibility Replication

Nov-09 : In this lesson, we will reanalyze/replicate COVID-19 spatial accessibility.

The purpose of this workshop is to reproduce Kang et al (2021) and in doing so, learn about methods for spatial accessibility analysis.

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.

Procedure

• identify an error, uncertain decision, or opportunity for improving reproducibility or study design. There’s a majority of folks interested in two feasible updates to this study, so let’s have everyone try both.
  • improve speed limit information. To do this, replace the network_setting function with use of the osmnx.speed module. Note how this changes the speed limit data checks before & after network_setting.
  • improve translation from hospital catchments into hexagons. To do this, focus on the overlap_calc function. Try commenting every line of this function with its purpose before changing anything.
• open the 03-COVID-19Acc-Reanalysis notebook and work with this notebook for this reanalysis.
• write a new markdown cell Original Study Design with one paragraph summarizing the original study, as the first section after the title markdown cell.
• write a new markdown section Original Study Design with one paragraph summarizing the original study
  1. description of your planned modification(s) with rationale(s)
  2. plan for visualizing/comparing/interpreting of results
  3. save as docs/reports/<name>-analysis-plan.md, e.g. docs/reports/holler-analysis-plan.md
  4. commit the plan and push to GitHub
  5. this is a small-scale version of a preregistration
• modify the main methods cell block(s) to note where you will make changes
• commit and push the changes
• implement the changes to code and save resulting map figures to the results folder before and after making each change, so that you can compare whether this changes results.
• interpret the results of your change(s) in a results & discussion section and discuss geographic threats to validity in this study.
• conclude with a statement about how the insights from this reproduction and reanalysis changes how we should interpret the original study results.
• save a new study report html form.
• enable GitHub Pages on the repository to serve the html version of the report to the web.
• write a blog post highlighting what you have learned from and contributed to with this reproduction and reanalysis study. In particular, please address these questions:
  • Keep in mind, a blog post is for highlights : aim for a post to be a short and impactful advertisement for the full study… no more than two paragraphs.
  • What have you learned from doing a reproduction study?
  • Has the reproduction deviated from the original study in any way?
  • If so, are the deviations improvements or errors?
  • Are there opportunities to further improve the study’s research design, reproducibility, or reproducibility for teaching purposes?
  • Are there opportunities to design meaningful replication studies to further test any theories established by this study?
• link to your full study HTML report and GitHub repository from the blog post

Note on Style: Remember that the primary motivation for reproduction and replication studies is not punitive. Frame your discussions in this report and previous reports in the constructive motivation for improving scientific knowledge through peer review. Project like CyberGISX generally, and the Kang et al 2020 publication specifically, are very new in geography, and our engagement with them should be both encouraging and constructive while emphasizing the value of open science.

Timeline

• Thursday, November 9 Class: Draft analysis plan of proposed changes and plans for interpreting the changes
• Thursday, November 9 Lab: Finalize analysis plan and commit the plan to GitHub
• Before Thursday, November 15 Lab: Commit finalized analysis report and blog post to GitHub

Main Page