Are we visiting greenspace differently during the COVID-19 pandemic? An innovative method to measure greenspace use by using mobile phone app data
Engagement with natural areas has increased during the COVID-19 pandemic, which may well form one of the enduring legacies of this time.
A better understanding of human interactions in urban greenspace, and how these have changed, is critical for decision makers to adequately manage and direct resources within these natural spaces. There is also a wider implication on public health as we recover from the pandemic.
The emergence of big data from the use of mobile phones provides researchers, city planners, and policy makers an opportunity to understand the use of natural spaces at a high resolution in a local setting. This data offers an innovative approach to doing research, with the ability to ask important behavioural questions retroactively and inform management decisions in a way that traditional forms of data collection (survey, satellite images, or GIS dataset) find difficult to achieve.
The importance of urban nature during the pandemic
The human-nature dynamic is important for our cities in normal times but during a pandemic, the perceived benefits of natural spaces are amplified, with greenspace playing an even greater role in promoting the health and wellbeing of our urban societies. Nature has been a source of physical and mental solitude for many, with lockdown rules heightening our appreciation for local parks and greenspaces. People are spending more time in nature than before the pandemic, however, access to nature is not always equal. People are less likely to have visited a natural space if they are living in an area of high deprivation, have a low income/education level, or are not working. The limited use of public transportation also limits the potential for green space usage during the pandemic. Furthermore, people who work from home were more likely to access outdoor natural spaces than those who commute to work, and people living closer to their nearest public green space are more likely to visit than those living further away.
Mobile phone data can change the way we measure greenspace use
Traditionally, the survey/questionnaire is the most common tool used to understand interaction with urban greenspace. While fundamental to understanding the wide-scale changes in preferences and social norms towards nature spaces, the broad focus and sampling strategy of primary survey-based techniques cannot provide the near real-time insights needed at a local level to inform management strategies. The global penetration of smartphones and the integration of Global Positioning System (GPS) technology mean near-real-time dynamic observations on unprecedented scales are now within reach. Mobile phones generate large volumes of opportunistic behavioural data, which provide the possibility to expand our understanding not only of where people are and what they do but also what they value. This can be applied to shine a light on the use of urban greenspaces throughout the pandemic.
What are the key questions that practitioners want to answer?
Researchers at the Urban Big Data Centre recently connected with the Park Managers Forum and stakeholders engaged with greenspace across Scotland to discuss the importance of data when making decisions. Stakeholders at the meeting were asked “what is the first challenge you would address using mobile phone data in greenspace?’
From the responses came two key areas of interest:
- To evaluate if greenspaces, parks, and playparks are functioning as expected.
- To inform decision making on how to best plan and disseminate resources on greenspace.
Mobile phone data presents an opportunity to address these points by providing near real-time behavioural data on how people use natural spaces and which spaces are in high or low demand. For example, we can study when and how people access and leave greenspaces, how long they spend there, where they come from and how these dynamics have changed over time. Understanding these aspects of greenspace use then can allow managers to address the second key challenge by making adaptations to greenspaces such as adding benches and playing facilities in a park, improving accessibility, and removing unnecessary facilities.
Mobile phone app data in Glasgow’s open spaces
Using 2019 and 2020 mobile phone app data from provider Huq, we extracted all GPS points (impressions) within a selection of Glasgow’s open spaces. We extracted over ¬769,000 impressions across two years. The sites with the most impressions were Kelvingrove Park, Glasgow Green and Pollock Country Park (Figure 1 and Table 1).
Figure 1: Huq impressions within Glasgow’s open spaces (2019/2020)
Table 1: Top 10 open spaces in terms of Huq impressions (2019/2020)
|1||Kelvingrove Park||West End||55497||1614|
|3||Pollok Country Park||Cardonald||42299||300|
|4||Small community park||Castlemilk||38038||15563|
|5||Small community park||Parkhead||27499||53195|
|6||Queens Park||Ibrox & Govanhill||20142||446|
|7||Botanic Gardens||West End||18731||1164|
|8||Hogganfield Park||Carntyne & Easterhouse||16958||445|
|9||George Square||City Centre||15143||14146|
|10||Small community park||City Centre||13750||11115|
Site-specific example: Alexandra Park before and during the Covid-19 Pandemic
For one site of specific interest to policy makers, Alexandra Park (Figure 2), we tested some more detailed analysis in terms of the changes in use caused by the COVID-19 pandemic.
Figure 2: Location of Alexandra Park
Analysis shows some changes in the morning peak at the park before and during the pandemic (Figure 3), with some differences in the daily number of impressions between the periods.
Figure 3: difference in impressions in Alexandra Park before and during COVID-19
The introduction and lifting of various lockdown rules are reflected in the visitation pattern throughout 2020 (Figure 4). After the first lockdown was imposed in March 2020 there is a large drop in visits to the park which did not recover until the end of 2020, before a 2nd national lockdown appears to reduce numbers again.
Figure 4: visits to Alexandra Park during 2020
In terms of who visits the park, Figure 5 shows the distribution of visitors based on their home region. We can clearly see that the surrounding area makes up a large portion of those visiting Alexandra Park.
Figure 5: where do visitors to Alexandra Park come from
While the health and well-being benefits of greenspace access have been increasingly recognised, they have taken on even greater significance over the last 16 months due to the COVID-19 restrictions. These same restrictions may also have widened inequalities in access, notably through restrictions on the use of public transport which is particularly important for the mobility of lower-income groups. Following on from the initial findings shown above, this ongoing research will assess how new forms of spatial big data can be used to better understand human interactions with urban natural spaces, through the case study of the Glasgow City Region, including the impact that the COVID-19 pandemic has had on inequalities of use. We also seek to explore the inherent biases of mobile phone app data and develop correction techniques, which achieve more representative population coverage for mobility research in general.
Notes and limitations of the analysis
The analysis presented here relies on the number of impressions (GPS points) and visits to open spaces in the City of Glasgow, where multiple impressions in a single day are treated as one visit. Only impressions with accuracy better than 100m were included. The analysis does not account for GPS points that may fall inside or outside of the open space boundary in error due to the limited accuracy attributed to the GPS point. Future research will overcome this by relying on stop detection techniques which looks at a user’s impression in time sequence to ensure those passing by the open space are not mistakenly included as part of the analysis. Because the number of active users in the Huq dataset is changing over time, we must normalise the data for analysis. The results presented in figure 3 are aggregated by hour on a daily basis. The results of figure 4 are normalised based on the total number of users in the Huq dataset per day during the period of analysis. For this reason, the absolute number of visits is not meaningful. The results of Figure 5 are based on the estimated home datazone of visitors and the map is clipped for the purpose of visualisation.
Mobile phone and social media data for human-nature research - 22 July 2021, 14:00 - 16:30 (BST)
This session aims to provide a venue for studies that use social media and mobile phone data for the assessment of human-nature interactions and preferences towards the natural environment.
Michael is a Research Assistant at the Urban Big Data Centre where he works on human mobility through spatial big data. Alongside his role in UBDC, he is also an environmental consultant and completing his PhD in Natural Resources and Environmental Management.