Reproducible Gis Science

The development of GIS software in an open source environment and the practice of using GIS to conduct open source science both have the potential to address the reproducibility crisis in geography. Open source software development ensures that researchers can use the same software used in a particular study without having to pay for proprietary software or be in a certain physical location to use it. (S. J. Rey, 2009). This makes conducting reproduction studies, where the exact data and methods used by the original study are used in the new study, easier and more accessible (NASEM, 2019. Reproduction studies also benefit from the use of open source GIS in source studies. The principles of open source science dictate that the exact data and code used in a study be publicly available. Thus, when researchers do open science in addition to using open source software, they make sure that their data and code are available to the public for verification and reproduction - not just the tools they used. Furthermore, when researchers conduct their studies using the principles of open science, they pave the way for future researchers to corroborate, challenge, and refine their conclusions through replication studies, in which the question of the original study is asked in a new context or altered slightly in the original context (NASEM, 2019). It’s important for researchers attempting to replicate a study to have access to the original data and source code so that they can focus on improving or adopting the study, rather than reconstructing it. Even beyond making replication studies more efficient, open-source science makes it possible for researchers to learn from eachother’s tools, strategies, and disciplines to ask new questions, or perhaps address old ones in new ways.

In addition to making research more rigorous, open-source GIS has the potential to contribute to positive cultural change in the research community. Part of the crisis of reproducibility within geography stems from the immense publishing pressure imposed on researchers by their academic institutions. Their career advancement is not determined by the quality of their work, nor the depth of their contributions to the scientific community, but by the number of papers they publish and how flashy their results are. This creates an incentive for some researchers to falsify or contort their conclusions to improve their chances of getting published. By encouraging transparency in data and code, an open-science framework would make it very difficult to publish fake or altered results (Ainsworth, 2019).

Open-science also has the power to relieve some of the publishing pressure on researchers. Open-science is all about making knowledge available to others, so in an open-science context, it doesn’t make sense to lock up peer-reviewed studies behind corporate paywalls. If academic publishing firms lose their monopoly on new research, universities and other instituions will undoubtably need to find new ways to evaluate researchers that don’t force them into rutheless competition with their peers (Ainsworth, 2019). While open-source practices would make it easier for reseachers to reproduce the findings of their peers, refine existing knowledge through replication, and reduce the toxic publishing pressure for academics, it cannot solve the reproducibility crisis on its own.

Even though open-source science would take power away from publishing firms, their absence does not mean that academia will cease to value shocking new discoveries over more humble, step-wise gains. Nor does it mean that exclusivity in the field won’t find a new way to manifest. In order for the benefits of open-science to reach their full potential, the scientific community will need to reassess its priorities and how they are expressed, think critically about barriers to accessing science and becoming a scientist, and reshape its relationship to the commercial market.

Other areas in which open-source GIS falls short include updates and documentation. The decentralized, volunteer based development of open-source software means that it is constantly evolving and not always well documented along the way. This can hinder efforts to reproduce and replicate existing research because the exact version of the software used in the original study may no longer be available, and the differences between versions hard to discern.

Just because open-source GIS can’t fix the reproducibility crisis alone, it doesn’t mean it’s not worth adopting. After all, open-science is based on the idea that the best, most innovative solutions are those that build off the work of others. Open-source GIS itself can thus be a stepping stone in the conversation about the crisis of reproducibility. Maybe someone will come along and add to it.

Based on the following readings:

• 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. http://dx.doi.org/10.1007/s10109-009-0086-8
• Dr. Rachel Ainsworth discusses open science culture in 2019: https://youtu.be/c-bemNZ-IqA