Fusing machine learning with place-based survey methods: revisiting questions surrounding perceptual regions

Abstract

This article explores questions on perceptions of the location of the ‘Midwest’, a contested vernacular region of the United States. We created a custom online survey with R’s web framework Shiny, in which participants were presented with a blank web map and asked to ‘draw’ their definition of the Midwest. Instead of simply describing the aggregated results, we employ machine learning algorithms – Naive Bayes, Random Forest and Categorical Boosting – in an attempt to classify users into groups, with a focus on the features that most effectively separate responses. We also demonstrate a way to engineer features from a single spatial response question and provide an implementation through a small R package. Furthermore, we discuss misclassified observations and suggest some driving factors in the construction of regional perception. This research is important not only for its contribution to perceptual regions but also for the approach, which could be applied to place-based survey analysis more broadly.

Keywords:

• Machine learning
• feature engineering
• perceptual regions
• vernacular regions
• the Midwest
• spatial cognition

Acknowledgments

The authors would like to thank Stefan Kuethe for extending the functionality of openlayers to meet the project’s needs. They would also like to thank Madeline Lundquist for georeferencing and digitizing the map reproductions in Figure 1 and Jonathan Rylander for helpful comments and edits on the manuscript. Additionally, the authors are grateful to the students who completed the survey. Finally, the authors would like to thank the anonymous reviewers for their helpful comments and suggestions.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data and codes availability statement

• Repository for sfe: https://gitlab.com/mapsurvey/sfe.

• Data: https://gitlab.com/mapsurvey/perceptual-regions/-/tree/main/data.

• Analysis: https://gitlab.com/mapsurvey/perceptual-regions.

Additional information

Funding

This research was supported by the Office of Research and Sponsored Programs at the University of Wisconsin - Eau Claire.

Notes on contributors

M. Haffner

M. Haffner is an Assistant Professor of Geography at the University of Wisconsin - Eau Claire where he started in 2018. He earned his PhD in Geography from Oklahoma State University. He was the project lead and primary contributor to the conceptual framework, survey design, survey implementation, software development and analysis.

P. Hagge

P. Hagge is an Associate Professor of Geography at Arkansas Tech University where he has taught since 2012. Prior to teaching at Arkansas Tech, he earned his PhD. in Geography from Penn State University. He contributed to the survey design, survey implementation and analysis.

C. Brown

C. Brown graduated from the University of Wisconsin - Eau Claire in May of 2020 with a Bachelor’s of Science in Geospatial Analysis and Technology. He is currently pursuing a Land Surveying and Geomatics Certificate at Western Colorado Community College. He contributed to the survey design, software development, data profiling and visualizations.

R. Heyrman

R. Heyrman graduated from the University of Wisconsin - Eau Claire in May 2021 with a Bachelor’s of Science in Geography with a minor in Geology. He is currently a GIS Analyst at McMahon Architects/Engineers. He contributed to the survey design, software development, data profiling and visualizations.

C. Perkins

C. Perkins graduated from the University of Wisconsin - Eau Claire in December 2019 with a Bachelor’s of Science in Environmental Geography. He is currently a Design Engineer at CCI Systems. He contributed to the conceptual framework and analysis.