Reproducible and Open GIScience

Sep-19 : In this lesson, we will debate the roles of open source GIS and Open GIScience.

Reading

• NASEM. 2019. Reproducibility and Replicability in Science. Washington, D.C.: National Academies Press. DOI: 10.17226/25303
  • Chapter 3, Understanding reproducibility and replicability (pages 31-43 )
• Rey, S. J. 2009. Show me the code: Spatial analysis and open source. Journal of Geographical Systems 11 (2):191–207. DOI: 10.1007/s10109-009-0086-8
• Dr. Rachel Ainsworth discusses open science culture: https://youtu.be/c-bemNZ-IqA

Discussion

In preparation for this debate, please prepare a reflection/position statement on the two questions in the form of a blog-style post on your GitHub site. Specifically, you might reflect on:

1. To what extent does open source GIS help solve the problems of the reproducibility crisis for geography? How?
2. Are there problems with reproducibility and replicability in geography that open source GIS cannot help solve?

While you read these, consider:

• What are the meanings of free software, open source, and open science?
• How is free as in the free software movement different from free as in a cost of $0.00?
• In what way might something be open source but not free?
• As we practice Open GIScience in this class, what are the implications for:
  • Learning
  • The Honor Code
  • Evaluation & Grades
• Is there space for Open GIScience in an undergraduate liberal arts education?
• What risks/benefits do you see for use of open source GIS in:
  • science?
  • government?
  • private businesses?

We can debate this in the context of this ESRI tool for managing COVID testing sites

Resources

• High-profile coronavirus retractions raise concerns about data oversight, Nature
• NSF Project to create a geospatial virtual data enclave and Doug Richardson’s short article about it as an alternative to data centers like Rocky Mountain Research Data Center
• Sergio Rey has been central in developing the Python spatial analysis library PySAL

Additional resources on Open GIScience

• Singleton, A. D., S. Spielman, and C. Brunsdon. 2016. Establishing a framework for Open Geographic Information Science. International Journal of Geographical Information Science 30 (8):1507–1521. http://dx.doi.org/10.1080/13658816.2015.1137579
• Sui, D., and P. Kedron. 2021. Reproducibility and Replicability in the Context of the Contested Identities of Geography. Annals of the American Association of Geographers 111 (5):1275–1283. DOI:10.1080/24694452.2020.1806024.
• Stodden, V. 2014. What computational scientists need to know about intellectual property law: A primer. In Implementing Reproducible Research, eds. V. Stodden, F. Leisch, and R. D. Peng, 325–340. Boca Raton: CRC Press.
• The Open Source Geospatial Foundation (OSGeo) take on open source
• The MIT License
• Creative Commons Licensing
• GNU’s Philosophy of Free Software and how open source misses the point
• How is OpenStreetMap having a moment concurrently with the death of an open source business model?

Landscape of open standards and software

• The Open Geospatial Consortium (OGC) develops standards, including simple features standards
• The Open Source Geospatial Foundation (OSGEO) supports open source geospatial software and education
• Geometry Engine Open Source (GEOS) is an open source code library in C++ for implementing the OGC simple feature standards
• PROJ similarly implements open standards for map projection.
• Who uses GEOS and PROJ?
  • PostGIS does!
  • Shapely does, which is in turn used by GeoPandas for spatial data science in Python.
  • GDAL too! and everyone uses GDAL, including…
  • Simple Features for R, the basis for spatial analysis in R, uses GEOS, PROJ and GDAL.
  • as does QGIS!

A ‘Show Me the Code’ Example: Distance Matrix in QGIS

The Distance Matrix algorithm documentation says nothing about how it calculates distances!

Watch a video of the issue, and/or follow the notes below.

Reading code to understand an algorithm

• Go to the QGIS GitHub repository github.com/qgis and search for distancematrix within the repository
• distance matrix algorithm source code
  • line 179: Transforms the target features into the source CRS (coordinate reference system)
  • line 182: CRS set to source CRS
  • line 183: Ellipsoid set to context ellipsoid
• distance area class source code
  • line 413: converts to meters
  • the only information passed to qgisdistancearea is the context ellipsoid and transformation, and the source crs
• The API (application programming interface) guide can also be a quick and helpful reference to QGIS variables qgis.org/api
  • there is a context.ellipsoid() and a context.distanceUnit(), but the distance matrix algorithm never uses distanceUnit.

Conclusions

• Distance Matrix appears to calculate distance in meters, no matter what.
• Target features are automatically re-projected to match the CRS of input features
• If the input layer uses a projected coordinate system, that system is used or planar distance calculation.
• If the input layer uses a geographic coordinate system, the project ellipsoid is used for ellipsoidal distance calculation.

Main Page