Independent Project Ideas

Nov-14 : In this lesson, we will generate feasible ideas for an independent project.

During this lesson, we will discuss and decide upon project ideas. Projects should be similar in difficulty and scope to one of the four labs this semester, and should encompass a research study using spatial analysis with open source GIS technology. Projects should contribute to at least one of these goals:

• increasing transparency/reproducibility
• assessing the reproducibility of prior research
• increasing our knowledge about a theory / process through replication

Sign up for a Thursday, November 16 check in time

Fall 2023 Project Ideas

• Complete Spielman et al replication by analyzing the index stability over time (pre-analysis plan has been written)
• Complete Chakraborty replication with new data (pre-analysis plan and preliminary work has been done)
• Replace most parallel processing code in Kang et al with Pandas/Geopandas code as we did for the overlap analysis / complete SVI portion of the study
• Replicate social media studies of disaster as in https://github.com/GIS4DEV/OR-Dorian
• Reproduce wildlife corridor analysis in R or Python, which has been done in QGIS model builder with some SAGA tools
• Replicate data analysis of rainfall and/or river flow for Vermont 2023 storms, ideally starting with R scripts from the USGS https://owi.usgs.gov/R/ (has this site changed? alternative sources may be https://github.com/orgs/USGS-R/repositories or https://www.usgs.gov/mission-areas/water-resources/science/science-topics/data-science )
• Replicate any of the studies we have done in different geographic regions, for different time periods, or otherwise with different data
• Otherwise significantly modify the methods approach for any of the studies we have done.
• Treat tutorial or lab problems from introductory GIS as studies to reproduce, and create R or Python code to solve the problem. Middlebury students can view the course manual and video tutorials. Let me know if you need access to particular sets of input data.
• Reproduce and update our prior studies on spatial analysis of COVID-19. These were written in seperate scripts, and not in our modernized compendium template. They use sptial statistical methods.https://github.com/HEGSRR/RPr-Saffary-2020, https://github.com/HEGSRR/RPr-Vijayan-2020, or https://github.com/HEGSRR/RPr-Mollalo-2020
• Reproduce any studies on CyberGISX / I-GUIDE / Hydroshare
• Propose another idea!

How to identify other project ideas

• Attempt to reproduce a straightforward published research study, especially one published with data and code, or with easily accessible public data sources. Reproduction will likely include some degree of reanalysis, in which some parameters of the study are intentionally changed to improve study design or facilitate learning. If you search for new papers, try adding terms like “R”, “Python” or “GitHub” to the search.
• CyberGISX and Hydroshare all contain published projects.
• Develop a project to learn a new package or technology for geospatial analysis or visualization.
• Replicate an existing study by substantially changing the geographic and/or temporal context

Requirements and Expectations

• The independent project will be due with the final GitHub site portfolio, during finals week.
• Projects may be completed in Python, R, or PostGIS SQL. Other languages or open GIS systems may be possible with approval.
• Students may choose to work on very similar projects and troubleshoot challenges with one another, but should maintain thier own complete version of the project and disclose collaboration through authorship and a paragraph in their report referencing CRediT roles
• Projects should include a reproducible research compendium in the form of a GitHub repository, paired with a report/blog post on your main GitHub site.
• You may ask the professor for feedback on the project up to Monday, December 11.

Procedure

• Start a new repository with the https://github.com/HEGSRR/HEGSRR-Template Use this template feature
• Customize the readme.md with project information
• Populate either the python notebook or Rmarkdown file with an analysis plan
• If your chosen study already has data available, populate the data folders with the available data
• Use the metadata_template.md template to create a metadata document for each data layer input and saved output
• Update the index tables data_index.csv and procedure_index.csv
• Commit and push the repository before you start running anlaysis
• Implement analysis in your Jupyter python notebook or Rmarkdown file
  • Document and manage the computational environment: in R use groundhog; or in Python use a Conda environment.yml file or referenc the CyberGIS system and kernel type
• Add results, discussion, and conclusion sections
• Render/save an .html version of the final study
• Link to the rendered report from the main readme.md
• Final round of checking and revising metadata documents, data_index.csv, procedure_index.csv and results_index.csv
• Short blog post highlighting the study and linking to the 1) repository and 2) html report. Seriously, just a paragraph or two is sufficient to highlight your contributions and link to the study.

Main Page