Menu Close

Research with app in Rembrandtpark published!

At last – the first real milestone in my PhD adventure! Here a brief description of the research published in our open access article: Schrammeijer, E. A., van Zanten, B. T. & Verburg, P. H. (2021) Whose park? Crowdsourcing citizen’s urban green space preferences to inform needs-based management decisions. Sustainable Cities and Society. 74, 103249

Parks, and other urban green spaces, are important in the city for many reasons – also for our health. They not only provide a refuge for animals, but they absorb excess rainfall and cool the city. Recently there has been a lot of discussion about how important green spaces are for mental health. But, as more and more people use green spaces it may be that they are less able to provide mental restoration. For this reason it is important to know how people experience different types of green spaces. Only then can public space be designed and managed in a way that makes them suitable for different needs. We know that places that are attractive, quiet, relaxing and safe are better for mental restoration. But what is attractive, and when is a place relaxing? These are all characteristics that are subjective – that different people can have different opinions about – and that are difficult to measure.

The Rembrandtpark

One way to measure how people use green spaces is by using ‘big data’ (eg. social media) and new mobile technologies (eg. apps on smart phones and watches). Many cities are investigating ways to become ‘smarter’ and incorporate digital infrastructure. These new ways of gathering information from people (aka crowdsourcing) are seen as promising. But, how well can these new technologies translate the way large numbers of different people feel about green spaces into useful information for urban planners?

To answer this question we took the Rembrandtpark in Amsterdam as a case study. We used a dedicated mobile phone app (The Mijn Park app) to collect information from people while they were in the park at certain locations. We compared this with other methods of gathering information, such as simply observing what people did, but also by downloading data from Flickr and the Vebeterdebuurt app.

We found was that using observations together with the dedicated app provided valuable information about places that were a priority for improvement. For example, places that felt unsafe but were still used by many people. The app also clearly indicated the most relaxing places. Social media (Flickr) uploads seemed to be more and indication of user quantity than use quality, and did not say anything about how relaxing the location was. Interestingly people did tend to complain via the Vebeterdebuurt app more about rubbish and other issues in places that were found to be more relaxing and attractive.

This research told us not only a lot about how people use and experience the Rembrandtpark, but also showed us how different types of information can give different insights. For example using only social media would give us a distorted view of which places are valuable in the park. Research has found that walking, relaxing and observing nature are the most important uses of green space – more important than active or social activities. But, active and social activities are much easier to observe or measure with big data and therefore they tend to overshadow the other essential uses of green spaces in the city. We have to be careful how we measure these more subjective social functions, especially as we move to a digitalised society and ‘smart cities’. A smart city ensures that it’s parks and public spaces can be used by all those who need them.

Maps of locations in the Rembrandtpark showing a) subjective quality results (Relaxing), b) objective quality results (Use quality), c) Social Media uploads and d) Verbeterdebuurt uploads. The numbers at each point indicate the study location number for identification. Some locations are missing in a) due to being filtered out with lower than 5 respondents. Each dot in c) and d) is a geo-located upload and the circles indicate the 50m buffer used around each study location. Image from publication.