The landscapes we interact with on a daily basis influence not only our behaviour but also our physical and mental well-being [1]. For example, providing access to “green lungs” in the form of urban parks, was already considered important in the 19th century. However, landscapes are not just static arrangements of physical objects, but rather the product of an interaction between humans and the environment. Our perception of landscape varies according to factors such as age, gender and cultural background. Analysing natural language has the potential to shed light on underlying perceptions and categorisations, but how can we find out how people talk about different landscapes and the language they use to describe such landscapes?
In July 2016 Pokémon GO was released. It had a major impact on acceptance of location-based games by the general public, and a considerable impact on human mobility patterns [2]. Suddenly, people walked through unknown streets and ventured into mysterious places, all for virtual rewards in a virtual world. What sparked my scientific interest was how Pokémon GO used volunteered geographic information (VGI) generated in a previous successful location-based game called INGRESS. Users would opt-in to upload a new point of interest (POIs) by providing a photograph of a “culturally significant” object or place and a short text description. Even though this process was decoupled from the actual gameplay, users generated millions of POIs, essentially providing a spatially heterogeneous (e.g. urban vs. rural) but in some areas such as urban centres very detailed map of objects they perceived to have cultural value. Coincidentally, at the beginning of the Pokémon GO phenomenon, I attended a colloquium talk on using simple games to validate land cover images and the potential of games to generate high quality spatial information at low cost [3]. I thought: since data contributed by a large number of non‐experts are increasingly used to validate and curate land cover data, couldn’t we use game elements and, in particular, the unprecedented interest in location-based games to make land cover data collection fun?
So, in my Master’s thesis, I set out to develop, implement and analyse a location-based game for in-situ land cover data generation. “StarBorn” was a location-based game with a strong focus on game play. Users conquered game‐tiles by visiting real‐world locations and collecting land cover data. Within three months, StarBorn generated 13’319 land cover classifications from 84 users. I was able to show that data are concentrated around users’ daily life spaces. As one might imagine, agreement was highest for urban and industrial land cover and user‐generated land cover classifications exhibit high agreement with an authoritative data set. However, I also observed low user retention rates and negative correlations between number of contributions and agreement rates. My master’s thesis won an award for outstanding scientific achievements at the Faculty of Mathematics and Natural Sciences of the University of Zurich and the results were published in the journal Transactions in GIS [4].
After my first successful venture into the domains of gamification and crowdsourcing I couldn’t shake the feeling that I had merely scratched the surface of the potential of using entertainment as a means of rewarding participation. In my current PhD project, I am revisiting the questions mentioned at the start of this article: how can data on how people talk about different landscapes and the language they use to describe such landscapes be collected? Are location-based games a viable methodological approach to generating high-quality natural language landscape descriptions? Can we build a corpus of how different people describe landscapes and can this corpus tell us something about how landscapes are perceived, categorised and communicated? And maybe most importantly, can we use fun to fuel science?
[1] Seresinhe, C., Preis, T. & Moat, H. Quantifying the Impact of Scenic Environments on Health. Sci Rep 5, 16899 (2015) doi:10.1038/srep16899
[2] Colley, A., Thebault-spieker, J., Lin, A. Y., Degraen, D., Häkkilä, J., Kuehl, K., … Schöning, J. (2017). The Geography of Pokémon GO: Beneficial and Problematic Effects on Places and Movement. Chi 2017. https://doi.org/10.1145/3025453.3025495
[3] Salk, C. F., Sturn, T., See, L., Fritz, S., & Perger, C. (2016). Assessing quality of volunteer crowdsourcing contributions: lessons from the Cropland Capture game. International Journal of Digital Earth, 9(4), 410–426. https://doi.org/10.1080/17538947.2015.1039609
[4] Baer, M., Wartmann, F. M., & Purves, R. S. (2019). StarBorn: Towards making in-situ land cover data generation fun with a location based game. Transactions in GIS, 1–21. https://doi.org/10.1111/tgis.12543