The activity tracking app ‘Strava’ was used by approximately 700,000 cyclists in the Netherlands in 2023; it is the largest activity tracker for cyclists. For most users, road cycling, mountain biking, or running are their main activities. However, once the app is installed, Strava also records many ‘commuting’ or utilitarian cycling trips. Since 2021, Strava has offered free access to the anonymized and aggregated spatial route data generated by its cycling and running users. Governments can gain access to the data of their district under strict privacy conditions and structures. More information can be found at (https://www.track-landscapes.com/privacybeleid-1 and https://www.strava.com/legal/privacy).
Strava cycling data is considered a highly valuable source of cycling data for regional/provincial cycling policies. In 2022, we compiled a collection of Strava analyses for the province of Utrecht, which showed the cycling route usage of sportive and utilitarian cyclists. In 2023, we continued the research with (1) additional insight into the representativeness of utilitarian cycling data in particular (see article), (2) a reflection on the regional cycling network and the ambitions of the cycling implementation program, and (3) an overview of the effects of cycling investments carried out between 2019 and 2023 on Strava cycling usage.
In this article, we provide an overview of the ways in which analyses of cycling flows are valuable for regional cycling policies. Both Strava sportive and utilitarian cycling activities were examined, as well as Strava (recreational) walking flows. The main focus of the research was on Strava utilitarian cycling data, as this represents everyday long-distance functional cycling traffic (see article 1 on representativeness), which is a core task of the Province. And since part 1 showed that Strava data is at least ‘reasonably representative’ for these movements, it is valuable to consider and incorporate these flows into cycling policies.
We now distinguish five approaches:
(1) Putting yourself in the movement
(2) Reflecting on different types of route networks
(3) Integrating the interests of various users on the cycling networks
(4) Visualizing and identifying missing links
(5) Moving across scales.
1. Putting Yourself in the Movement: A Means for Discussion, Reflection, and Justification of Upgrading Cycling Routes.
In both 2022 and 2023, work sessions were held with cycling experts from various organizations, focused on the Strava maps of utilitarian and sportive cycling flows. The maps resonate with professionals in the fields of mobility, spatial planning, and recreation because they provide a comprehensive picture of usage, based on actual cycling trips undertaken.
In the movement of people, there is experience; what do people know, experience, and perceive as routes with a certain continuity, recognizability, logic, or attractiveness? These are the human components that determine (Strava) cycling flows, and you can never fully capture them in a model. This does not mean that it is a replacement for traffic models, local counts, or other sources/forms of cycling data. Traditional count and traffic model data may be more complete and reliable in certain respects, but they are also more anonymous and therefore less engaging.
Strava data thus facilitates lively conversations and discussions about the logic and congestion of current cycling routes and the desirability of new or improved routes. The Strava maps allow for ‘immersion in the movement,’ which is valuable in itself. It is crucial, however, that you place yourself in the perspective of long-distance utilitarian cycling traffic (if using cycling ‘commute’ data) or road cyclists (if using cycling ‘leisure’ data), and not view it as general utilitarian or recreational cycling traffic.
It can then play a role in ‘justifying’ cycling investments. Based on the patterns of walking and cycling flows, you can reason (and thus justify) cycling experience, interests, logic, and the potential of the cycling network, and why a measure would be valuable. This can address both the upgrading of an existing route and advocating for a new (missing) route.
2. Reflecting on Different Types of Route Networks: Confirmation, Surprise, and Attention Through Broad Overview.
The strength of Strava Metro is that cycling flows of the two different types of cycling can be visualized separately on the scale of an entire province. The map below shows Strava utilitarian cycling flows on all paths and roads throughout the entire province. This map shows the utilitarian cycling movements. It displays the cycling flows of all paths and roads included in OpenStreetMap. The focus of usage lies in/from/to Utrecht, and then Amersfoort. Because this article is public, we need to express passages in 'percentages' (instead of absolute numbers). The legend is therefore expressed as a percentage of the most-passed path in the province (Biltsestraatweg). This path receives 'several tens of thousands' of passages per year. And in all work and residential centers, there are routes that see 'thousands' of passages per year.
In the documents and GIS maps we created for the province, the exact, absolute numbers are, of course, visible and available.
When looking at such a comprehensive overview, certain routes sometimes stand out that you hadn't noticed before or had forgotten about. In the work sessions, the usage of different routes was compared. Often, this matched the expectations/knowledge of the cycling experts, but some routes were labeled as having 'surprisingly high' or 'surprisingly low' usage. Not all heavily used routes are equally recognized as being of regional importance (for example, not yet included in the Regional Cycling Network or as a through-cycle route). Sometimes the usage was acknowledged/understood, but the route had simply fallen out of (policy) focus. The usage maps then serve well as a 'wake-up call.' On some routes, there was more surprise about the level of usage, which is a reason to further investigate the usage of those routes (e.g., through counts). In both cases, the Strava cycling maps can help to refine the importance of different routes.
However, the overall picture of usage (350,000 lines) is also somewhat overwhelming, making it difficult to distill insights. It helps to 'filter' or 'mask' the maps of cycling flows based on certain types of networks. For example, it is visually helpful to show only the roads that receive at least '5%' of utilitarian Strava passages per year.
Such an image is, for example, already easier to compare with the preliminary choices for through-cycle routes in the province; are the current or proposed through-cycle routes being used frequently?
But we can also reverse the question; are certain routes perhaps more important than we previously thought? We overlaid the proposed through-cycle routes as a white 'layer' over the Strava cycling flows. The remaining visible colored lines are therefore not (proposed) through-cycle routes, but they are still heavily used.
There are many routes that we have categorized into three types.
(1) Direct urban-connection routes; such as Koningsweg and Koeweg near Bunnik (1a), Hilversumsestraatweg between Baarn and Hilversum (1b), or between Amersfoort and Soest via Maleburg (1c).
(2) ‘Green alternatives’; routes with an attractive green character that are sometimes also the ‘shortest’ routes to residential centers, but perhaps their green character also has a positive impact on usage. Examples include (2a) routes along the Vecht between Breukelen and Maarssen/Utrecht, (2b) De Groenedijk and Lindenlaan between Utrecht and Hilversum, or (2c) routes through Korte and Lange Duinen as green alternatives to Birkstraat or Soesterbergsestraat between Soest and Amersfoort.
And (3) perhaps the most interesting type: ‘ring routes’. The current/proposed through-cycle routes are primarily radials towards the major centers, but total usage is a sum of cycling movements in various directions. Those who do not need to be directly in the central city often prefer to avoid it and follow an alternative through a less urban, greener environment.
Most cycling movements will partially make a more radial movement, but there are also many cycling flows that seem to make a more tangent ‘ring movement.’ Some examples: (3a) Above Bunnik, via Fort Vechten above Houten, over the Waijensedijk to Nieuwegein. And, (3b, the Amersfoort variant of this) around the industrial area ‘de Wieken,’ between Amersfoort and Leusden, to Treekerweg, Doornseweg, or Amersfoortseweg. This recognition is mainly an argument for the importance of a consistent and continuous cycling network in various directions, not only or primarily in the metropolitan radial directions. The tangent ring movement also encounters many barriers and missing connections. Most of the major infrastructure and large bodies of water have developed radially from/to the major cities (and vice versa). Those moving in a more perpendicular/ring direction are more likely to encounter this infrastructure.
Comparisons have been made in various other ways between Strava utilitarian cycling flows and certain types of networks:
Strava utilitarian cycling vs. the regional cycling network; which routes are not yet part of the RFN but are heavily used?
Strava utilitarian cycling vs. separated cycling paths; on which cycling paths is the congestion the greatest? How wide/narrow are these paths?
Strava utilitarian cycling vs. the recreational node network; aren’t the ‘recreational’ routes also heavily used for utilitarian purposes?
3. Integrating the interests of various ‘other’ users on the cycling networks: sportive cyclists, walkers, and runners.
The utilitarian cyclist is generally seen as the primary user of the cycling networks, but Strava data shows that there is also a lot of other types of movement on the cycling networks.
The use of routes by different moving people is partly different, but often also surprisingly overlapping. The sportive Strava cycling data provides the opportunity to advocate for the route interests and needs of an additional cycling group. Where can we consider the sportive cyclist in the design of cycling routes? And can we use sportive usage as an additional argument to improve routes?
But the biggest ‘eye-opener’ of the 2023 work session was the Strava walking data (recreational walking and running). This was seen as a very valuable addition; while there is some understanding of cycling flows based on counts and models, there is still a complete lack of insight into walking flows (for more info; see this article).
Strava shows that there is also a lot of recreational walking on specific ‘cycling’ paths and other roads. Where does this lead to conflict? Is a separate walking facility desirable? Or can we use the importance of walking to push missing links or the upgrade of routes higher up the agenda? This benefits both groups of users.
The maps below show walking and cycling flows in the Amersfoort-Woudenberg region, the focus area of the work session, with various discussed observations of heavily or lightly used routes indicated by numbers. For example, the Valleikanaal (1), routes around the Groene Zoom (2 & 3), and city-country connections on the border of Amersfoort (4).
4. Clear Visual Insight into Missing Links.
By looking at overall cycling flows, as well as walking flows, the missing connections become apparent. In the image above, several potential/discussed desired connections are indicated with dotted lines.
From the existing flows, it is easy to reason where a new connection would be a valuable addition. The reasoning could be that usage intensity does not continue smoothly between certain locations; there is visible evidence of people cycling a longer route. We prefer to extend lines where usage strongly converges as continuously as possible.
The diversity of users that you can separately uncover with Strava also helps in this regard—a new bridge, tunnel, or connection through rural areas is most valuable when it is of added value to various types of moving people. This increases the chances of getting it onto the implementation agenda.
5. Moving Across Scales
Where the 2022 work session primarily focused on Strava cycling flows across the entire province, the 2023 work session zoomed in on two sub-areas: (1) Utrecht-Zeist-Bilthoven and (2) Amersfoort-Leusden-Woudenberg. Cycling experts from those specific municipalities were present. Based on more local area knowledge and route knowledge, a much more targeted interpretation of the data can be provided.
For example, Strava cycling usage is compared against the perceived quality of the cycling routes: width, surface, continuity, environment, safety, but also the intensity of use from personal experience or other counts.
Moreover, local cycling officials often already have cycling (and walking) measures on their wish list or cycling agenda that have not yet been definitively implemented. The data helps them better support the importance of these measures with factual arguments about cycling and walking flows. And overview maps of these real flows are simply visually clarifying and convincing.
This also leads to an interesting question, which applies to any form of data: should we use data solely as ‘objective’ input that is meant to correct subjective human desires? And thus, it is only valuable if it is ‘representative’ data. Or, are we okay with using data, regardless of representativeness, if the data result suits us? There is no clear or simple answer to this, in our view, but it is at least important to recognize and acknowledge the difference.
Additional and Refining
In our view, Strava data does not definitively determine which route is best chosen for a cycling investment/upgrade (e.g., into a through-cycle route). First, because current usage is often not the most important criterion for such choices, but rather the potential for future usage. Second, because Strava does not represent all types of cyclists or walkers.
However, for those who are aware of this, Strava data does provide good opportunities to reflect on, discuss, and justify cycling route potentials based on the interests of various targeted types of users of networks for moving people in cities and landscapes. This is also how it is used in the Province of Utrecht; it serves as an additional and refining form of analysis alongside various other forms of cycling data, inventories, and criteria for cycling investments.
And while that is already a lot, we continue to see new applications emerging from this usage data. One of these, which you will read about in article 3, is monitoring and evaluating cycling measures.
Comments