Active transportation and geographic context:
analyzing movement behavior with micro-geographic
data
Gerd Weitkamp1, Paul Plazier1, and Roya Shokoohi2 1University of Groningen, Groningen, Netherlands
{s.g.weitkamp,p.a.plazier}@rug.nl
2 Hanze University of Applied Sciences, Groningen, Netherlands r.shokoohi@pl.hanze.nl
Abstract. Stimulating active living and active transportation requires specific
conditions of public spaces. To clarify the relationship between active transportation behavior and the geographic context, we collected GPS data from cycling trips of 24 participants for two consecutive weeks. We compared these data with geographic datasets, and our preliminary findings show for both conventional bicycles and electric bicycles relationships between movement behavior and road types, built environment, and weather conditions. The application of highly detailed spatio-temporal geo-data enables a better understanding of the relationship between movement and the geographic context.
Keywords: Health geography, active transportation, cycling, micro-geographic
data, GPS-tracking.
1
Introduction
Active living is a way of life that integrates physical activity into daily routines, and therefore can improve people's health and wellbeing. Active transportation, such as cycling or walking, is a type of physical activity that can provide substantial health benefits [1]. To encourage active transportation, we have to understand movement behavior related to these physical activities. Movement is a process that occurs as a response to properties of the individual, and is highly influenced by the geographic context across multiple spatial and temporal scales. Individual properties and the geographical context determine characteristics of movement behavior, such as speed and direction [2].
To clarify the relationship between active transportation and the geographic context, methodological refinements are needed. The availability of geographic- and movement data at various spatio-temporal scales enables to study movement behavior
at a high level of detail. Findings based on these data may be used for evidence-based design that can improve active living environments.
In this research we aim to better understand the relationship between cycling behavior and the geographic context, as it is the dominant mode of active transportation in the Netherlands [3]. Cycling behavior can be strongly influenced by many geographical conditions, more than other modes of transportation [4]. These geographic conditions may however have a different impact when using an electric bicycle.
2
Methods
We collected movement data with GPS tracking, and we used existing geographic datasets that included the built environment, road types, and weather conditions.
Movement data was collected from 24 participants, (12 men and 12 women, age 25-62), who lived and worked in the north of the Netherlands and who used an electric bicycle to commute from home to work. GPS data was collected throughout the day for two consecutive weeks between november 2015 and april 2016. We used QStarz BT-Q1000XT GPS tracking devices, and tracked the location of each participant at a 10-seconds interval. The spatial accuracy was 5 meters. After the tracking period, errors were removed from the GPS data, and the data were analyzed with V-analytics software, resulting in a point dataset with only trajectories that represent movement between places. The type of origins and destinations of places as well as modes of transportation were clarified with the participants during an interview after the GPS data was analyzed, and these were added to the point dataset. Only the trajectories with (electric) bicycles as a mode of transportation were used for analysis. The 24 participants recorded a total of 1090 trips, from which 66 by regular bicycle and 376 by electric bicycle.
Road type data were derived from a topographic dataset, the BRT TOP10NL from the Dutch Kadaster. It included information on type of user (e.g. cycling only or mixed use), and road width. Urban and non-urban areas were also derived from the BRT TOP10NL from the Dutch Kadaster. Weather data were available from a weather station at Eelde, located in the north of the Netherlands. The dataset includes daily averages for temperature, wind and rainfall.
Movement characteristics (e.g. speed) were analyzed in relation to the purpose of the trip, road types, built environment, and weather conditions. The statistical package SPSS was used for data analysis.
3
Results
Preliminary findings show a number of relationships between movement characteristics and geographic context. When comparing the movement speed to work-related with non-work related destinations, there was a significant effect for purpose of the trips, with work-related trips receiving higher scores than non-work related trips.
When comparing the speed in urban areas with non-urban areas, there was a significant effect, with the urban areas receiving lower scores than non-urban areas. In urban areas the speed may be lower due to the larger number of crossings, traffic lights and other road users. On the other hand, weather influences such as wind may have a stronger negative effect on the speed in non-urban areas.
We calculated the difference between wind direction and route direction. We found a correlation between this difference and speed, r = 0.25, p < .05. When wind direction and route direction where more similar, speed was higher. Since most participants used electric bicycles, it was not expected that the wind would be strongly related to speed. The temperature and speed were negatively correlated, r = -0.17, p < .05. Again, results may be different in urban and non-urban areas.
Speed differences were also noticed on different road types. A paired samples t-test showed statistically significant speed differences between separate cycling roads and roads with mixed use, with separate cycling roads receiving higher scores (M = 18.41) than mixed use roads (M = 14.27). When comparing speed on narrow bicycle lanes (width <2 meters) with wider bicycle lanes (width 2-4 meters), there was a significant effect for road width. Surprisingly, the narrow roads receiving higher speed scores (M = 19.24) than wider roads (M = 18.24). A possible explanation is that the classification of cycling roads is slightly different in urban and non-urban areas.
3
Conclusions and Discussion
This study used spatio-temporal data to discover relationships between movement behavior and geographic context across various scales. The findings show that participants change their movement behavior whith varying destination types, urban context, and weather conditions. In addition, some road types afford a higher cycling speed than others. These results can be used for designing spaces that encourage active transportation.
However, current results only show preliminary findings. Further research is needed to analyze and discuss more relationships. Moreover, the findings are based on the movement data of a small number of participants with a particular profile. A larger and more representative group of participants may show different results. The quality
of the geographic datasets could also be improved. Higher positional accuracy of GPS, higher temporal precision of weather data, and better attribute accuracy of topographic data is expected to enable better understanding of the relationship between movement behavior and geographic context. A future step would be to develop a model to predict the potential of active transportation in a given geographic context.
Acknowledgements. We thank Leon van der Meulen, junior GIS specialist at the Geoservice of the University of Groningen, for his support preparing the GIS datasets.
References
1. Sallis, J.F., Frank, L.D., Saelens, B.E., Kraft, M.K.: Active transportation and physical activity: opportunities for collaboration on transportation and public health research. Transportation Research Part A. 38, 249–268 (2004)
2. Dodge, S.: From Observation to Prediction: The Trajectory of Movement Research in GIScience. In: Onsrud, H., Kuhn, W. (eds.) Advancing Geographic Information Science: The Past and Next Twenty Years. GSDI Association Press. pp. 123–136. Needham, USA (2016)
3. KIM: Fietsen en lopen: de smeerolie van onze mobiliteit. Kennisinstituut voor mobiliteitsbeleid, Den Haag (2015)
4. Winters, M., Davidson, G., Kao, D., Teschke, K.: Motivators and deterrents of bicycling: comparing influences on decisions to ride. Transportation. 38, 153–168 (2011)
