Graduation work by Thijs Dolders & Mart Reiling
In this thesis we try to find ways to understand how people are using public space as a sports field by analysing 110.000 running activities from mobile running applications Strava and Runkeeper. In combination with the results of a survey, it forms the basis on which designs are made.
Where and when do people run? Physical activity in combination with landscape architecture is a field yet to discover, although more and more people are sporting individually, using public space as their sports facility. In Amsterdam the amount of runners doubled to 200.000 in between 2009 and 2014. Yet no knowledge exists on their active outdoor behaviour: where do they go? And what are they looking for in these places? In the aim to design ‘the runner friendly city’ we used a newly available method by counting great quantities at once: crowdsourced data.
The data has been present for some years now, but its applied value is still not researched seriously. The data of the 90.000 running activities in between February 2014 and February 2015 of the app ‘Strava’ and the 20.000 publically available running activities of the last five years of the app ‘Runkeeper’ are analysed to retrieve an image of the spatial behaviour of runners. This data is visualised in both graphics (Figure 1) for example to show the most run distances, the behaviour of runners in time and the (mapped) spatial distribution of running routes (Figure 2 & 3). This data is combined with an analysis of the spatial structure of the city, from where the relations between the configuration of space and spatial behavioural patterns can be clearly recognized.
Figure 1: Frequency of running distances
Figure 2: Runkeeper activities 4.5-9 km
Figure 3: Runkeeper activities >9km
From this data it becomes clear that people mostly run in parks and along water. However, some places remain relatively ‘underused’ in this overall image. Mapping only the short distance activities clearly illustrates the clear and small laps that can be found in the city, whereas long distance activities tend to use the ‘green finger structure’ of Amsterdam as a way to leave the city. When it is dark outside, people run closer to the city and stay away from places like the ‘Amsterdamse bos’.
What are the spatial requirements of runners? 54 surveys with runners in Amsterdam aimed to get a better understanding of the spatial needs that made people run in these specific areas. Many different spatial aspects were found that explained the behavioural patterns.
Runners’ needs were visualised in two ways. First, participants were asked to assign importance (on a 1-5 scale) to a list of spatial aspects relating to running. Second, they were challenged to draw their (last performed) running routes and phrase the positive and negative spatial experiences which they encountered during their run, which is made visible in a geotech wordcloud (figure 4).
Figure 4: Users experience
Rembrandtpark The Rembrandtpark was recognized in the data as one of the lesser used ‘large’ (>30 ha) parks of Amsterdam, although people pointed out these large parks were places where they prefer to run. Taking in regard the central location of this park, higher usage intensities could be expected here. Especially the west side of the park was very underused. During darkness, people completely avoided this western park-side (figure 5). Also, long distance activities did not pass the west, while short distance activities did (figure 2&3). The surveys explained this behaviour; runners indeed experienced the west to be unsafe in the evening and the park was considered to be ‘too small’. On the contrary the only slightly bigger Vondelpark was found to have a safe evening atmosphere and a ‘nice distance’.
Figure 5: Strava data during the dark (15.460 activities) zoom in on Rembrandtpark
The strategy of the park is to connect the west-side more clearly with its northern and southern bordering roads. This path is wide and available for both bicyclists and runners, creating an ongoing flow of people through the park, also in the evening. It simultaneously creates a clearer, enlarged main lap through the park. The strategy is further developed in two models of which the model ‘a false start’ makes a new lap close to the apartment buildings with ‘functions’ like an urban gym attached and leaves the middle of the park more forested and natural (figure 6). Model two ‘what goes around’ extents the ponds in the middle of the park in which islands are the place for ‘functions’, the new path borders the water (figure 6). This makes the middle of the park more open, visually connecting both sides (figure 7). The forested area close to the apartments remains. New paths form logical laps with specific distances and are smoothly connected.
Figure 6: From left to right ‘Existing situation’, ‘A false start’ and ‘What goes around’
Figure 7: Visualisation ‘What goes around’ in the light and dark
Schinkel The Schinkel is an old canal in the West of Amsterdam, connecting the ‘IJ’ with the ‘Nieuwe meer’. Although it is still a busy boating route, the embankments are not as busy anymore as they have been throughout history. Adjacent to the water, a lot of historic buildings remain, but also cafés and sport facilities. In the data this route can be seen as a missing link in the Amsterdam running network. People run next to big roads like the Amstelveenseweg instead of the quiet canal, which is located parallel to it. Running should be beneficial for health, however a busy road is not the best place to run. Where the Amstel as a similar water structure is used a lot to run from city centre to countryside, as well as to run short rounds, the Schinkel is only used in some short laps. Zooming in on the profile of the embankments, it becomes clear that not much space is available for runners, as most space is reserved for parking (figure 8). There are also some major perpendicular roads crossing the Schinkel which are obstructing the walk along the water (figure 8). Therefore there are two strategies needed: get rid of the parked cars by placing parking facilities every 500 metres (5 minute walking distance) on vacant plots and two solving the perpendicular roads by making tunnels underneath one side of the bridges.
Figure 8: On the right the map with possible spots for parking garages and on the left an inventorisation from huge perpendicular roads and current lay-out
With less parked cars there is room for bicycles and runners. All three users tracks (runners, bicyclists and pedestrians) get their own track (as the biggest nuisance was disturbance by other users) and get different surface materials, but all are coloured red (figure 9).
Figure 9: Concept running track in the city and lay-out boulevard
In this way the entire route becomes a recognisable structure; a red carpet or running track laid out for slow traffic. Where perpendicular roads cross the Schinkel, a possibility to go underneath the perpendicular roads by a tunnel is proposed (figure 10). The important connection point between Vondelpark and West of the city which is called Schinkelhaven, becomes a special place, where people can do exercises in an urban gym (20% off runners combined their run with strength exercises) and have the option to run to the Vondelpark or continue their route along the Schinkel (figure 11).
Figure 10: Red carpet for slow traffic
Figure 11: Schinkelhaven, a crucial square dominated by the urban gym
Conclusions The data forms a hugely important source of information, as it allowed to detect places which have serious structural problems with consequences for usage. Apart from the two elaborated cases, many more concerning patterns were detected through visualising the data. Not only small interventions can be based on small errors found through analysis of this data, but also errors on a large city scale. However the question remains who the data is representing. Though, finding bigger sources of information than these, is basically not possible. The data on itself is already interesting, but especially the visualisation, interpretation and translation of this data is the added value. Still, as spatial conditions are different in every city, design principles cannot simply be copied to another context.
The study also showed that designing for runners is never only beneficial for runners: it immediately showed errors for pedestrians and cyclists as well. These were solved simultaneously when designing for runners.
At last, an essential value of the data lies within its evaluation possibilities. Every year, new data becomes present, and therefore data-analysis, both before and after interventions, makes it possible to determine the effects of design interventions on usage.