VGIscience

Environmental volunteered geographic information for personal exposure awareness and healthy mobility behaviour (ExpoAware)

  • Carolin Helbig
    Carolin Helbig
    Helmholtz Centre for Environmental Research GmbH - UFZ
  • Uwe Schlink
    Uwe Schlink
    Helmholtz Centre UFZ
  • Torsten Masson
    Torsten Masson
    University of Leipzig
  • Anna Becker
    Anna Becker
    Universität Leipzig
  • Abdelrhman Mohamdeen
ExpoAware

Urban spaces are hotspots of environmental pollution such as noise, air pollution and heat. These stressors affect our health and are highly contextually distributed in space and time. Portable sensors have advanced the measurement of these stressors and allow mobile exposure measurements.

Environmental-related Volunteered Geographic Information (VGI), which is actively collected by non-experts, can thus simultaneously provide valuable knowledge about temporal-spatial exposures and the mobility behaviour of individuals. Smartphone-based detection methods also provide active options to capture users’ perceptions and opinions about exposure. This combination of subjective and objective VGI is a comprehensive data pool that not only allows exposure to be evaluated but also the behaviour and decision motives of individuals during everyday mobility.

In this project, we therefore work on three main questions:

  1. How can environmental VGI generated with smartphones be used to extract spatiotemporal movement patterns and exposures?
  2. How can we collect subjective data and analyse behavioural changes when confronting people with exposure data?
  3. How can environmental VGI be used to improve mobility and exposure models?

We use intelligent sensors and an interventional study design, based on a self-developed framework of contemporary geography and psychological theories, to collect VGI with individuals from different socio-economic contexts. We will analyse the data using GIS-based methods of space-time analysis and psychological test statistics. To compare empirical data with models, we perform agent-based simulations.

The main goal of this project is to explore the potential of environmental VGI to change mobility behaviour in line with a healthier pathway decision by raising awareness of exposure.

In the following we give an overview of the results of the project.

Mobile Sensors (Wearables)

City dwellers and participants in walking and cycling are exposed to environmental stressors. In our studies we used mobile sensors that can be worn by citizens to measure their exposure to environmental stressors. Within two measurement campaigns the cyclists and pedestrians have taken the sensors on their daily routes through the city of Leipzig.

The first measurement campaign took place from July 2020 to October 2020. A set of sensors was used here, consisting of the fine dust sensor Dylos DC1700 (PM2.5, PM10), the gas (NO2, NO, O3) and Leo/Ateknea temperature sensor, and a smartphone (noise, light, GPS, time stamp) (see photo). In the second measurement campaign from August 2021 to August 2022, the PAM (Personal Air Quality Monitor) sensor (PM1, PM2.5, PM10, NO2, NO, CO, O3, temperature, humidity, noise, GPS, time stamp) was used in combination with a mobile app (see below).


Figure 1: Sensor set of first measurement campaign


Figure 2: Sensor of second measurement campaign

Mobile App

For the second measurement campaign, a progressive web app (PWA) was developed. The technology was used because of its accessibility via any browser or easy installation on all smartphones as well as a relatively low development effort (compared to native apps). The app was used by the study participants to track the start and end of their routes as well as to collect answers about the route. In addition, participants were able to view their measured tracks after uploading the sensor data. Another feature of the app was that the participants were shown an alternative route, less exposed by environmental stressors, which they could then drive the next time.


Figure 3: PWA for tracking the routes of participants

Measurement Campaign

In two field experiments, we not only equipped cyclists and pedestrians with our sensor technology, but also used questionnaires to study their perceptions of environmental stressors as well as their intentions to change their routes in order to avoid pollution.

In both studies participants were allocated to a measurement group or a control group. The measurement group used the wearable sensors for three days and received an individualized feedback report at the end (study 1) or everyday feedback in a mobile app including alternative route suggestions (study 2). The control group did not carry the sensors or receive feedback, while both groups filled out multiple questionnaires throughout the study period. This procedure allowed us to identify causal effects of using the sensor kit and receiving feedback.

In the first study a final sample of 107 participants finished each questionnaire of our study. In the second study the final sample consisted of 137 participants.

In the questionnaires, we asked participants about psychological concepts, that can help to explain health behaviour. These predictors of healthy behaviour were based on Protection Motivation Theory (PMT, Rogers, 1975). PMT predicts healthy behaviour change, when threat perceptions (perceiving the current situation as a health hazard with severe consequences) as well as coping appraisals (capacity to adapt behaviours to lower the health hazard) are high.

Using a wearable sensor and receiving feedback on exposure levels makes health hazards such as particulate matter visible to the user. We tested, whether making these stressors visible, would affect participant’s threat perceptions and willingness to choose less polluted routes in the future.


Figure 4: PAM sensors prepared for the measurement campaign


Figure 5: Instruction of a participant in the sensor technology


Figure 6: Effect on Threat perceptions in study 1 and study 2

In both study 1 and study 2, we found that health threat perceptions for particulate matter increased in the measurement group, while they did not increase in the control group (Figure 6). This is an indication that wearable sensors have a psychological effect on the people wearing them.

Effects were less consistent when looking at participant’s intentions to change their routing behaviour. In general, effects of wearing the sensors had relatively short-termed effects on people’s intentions to take less polluted routes. Specifically, wearing the sensors increased intentions to take less polluted routes from pretest to posttest in study 1, albeit only for participants who had no strong habits for their routing choices at the start of the study (Figure 7).

In study 2 we found that participants in the measurement group increased their intentions to change their routing behaviour slightly more than those in the control group (Figure 8, marginally signifficant).


Figure 7: Effects on Intentions to change in study 1


Figure 8: Effects on Intentions to change in study 2

Visualisation Application

We implemented a prototype of a 3D visualization and analysis application with which the distribution of stressors (hot and cold spots) within the city can be analysed and evaluated. The application is based on the Unity game engine, a development environment for computer games and other interactive 3D graphics applications. Using Unity enables us to combine methods of Visual Analytics, 3D visualisation, and Virtual Reality as well as to implement analysis methods and make them available via a user interface. For further details on and the download of the application see Helbig et al. 2021.


Figure 9: 3D application for exploring and analysing the mobile sensor data

Publications

  1. Helbig, C., Ueberham, M., Becker, A. M., Marquart, H., & Schlink, U. (2021). Wearable Sensors for Human Environmental Exposure in Urban Settings [Journal Article]. Current Pollution Reports, 7. DOI: 10.1007/s40726-021-00186-4
  2. Schlink, U., & Ueberham, M. (2021). Perspectives of individual-worn sensors assessing personal environmental exposure [Journal Article]. Engineering, 7(3). DOI: 10.1016/j.eng.2020.07.023
  3. Becker, A. M., Marquart, H., Masson, T., Helbig, C., & Schlink, U. (2021). Impacts of personalized sensor feedback regarding exposure to environmental stressors [Journal Article]. Current Pollution Reports, 'in print'.
  4. Helbig, C., Becker, A. M., Masson, T., Mohamdeen, A., Sen, Ö. O., & Schlink, U. (2022). A game engine based application for visualising and analysing environmental spatiotemporal mobile sensor data in an urban context. Frontiers in Environmental Science, 10. DOI: 10.3389/fenvs.2022.952725
  5. Helbig, C., Becker, A. M., Masson, T., Mohamdeen, A., & Schlink, U. (2023). Individuelle gesundheitsrelevante Umweltexpositionen im Rad- und Fußverkehr – Trends, Auswirkungen und eine Fallstudie zu Resilienzs. In S. Kabisch, E. Banzhaf, & D. Rink (Eds.), Die Resiliente Stadt: Vol. 'in print'. Springer Nature Open-Access Fachbuch.
  6. Becker, A. M., Masson, T., Helbig, C., Mohamdeen, A., & Schlink, U. (2023). Wearable sensors increase environmental health threat perceptions: Results of a randomized field study: Vol. 'under review' [Journal Article]. 'under review'.