OSM Can-BICS (English)

Evidence strongly supports that  where all-ages-and-abilities bicycle infrastructure is common, there are higher rates of cycling and other forms of active transportation . The type and quality of bicycle infrastructure is important; for example, infrastructure that is away from traffic noise, with beautiful scenery, and physically separated from traffic is known to  motivate riding . These types of infrastructure may also increase diversity among people who ride bikes, for example, some bicycle infrastructure may help  reduce the gender gap . The  Canadian Public Health Officer’s 2017 Report  called for greater attention to the built environment to understand how community design can support healthy living. Further exploring of these topics depends on high-quality data about bicycle infrastructure in Canada.

However, data on bicycle infrastructure nationally are limited. To further advance knowledge around how supportive cities area for cycling, find change over time, and inform future infrastructure projects in a Canadian context, we need a detailed and meaningful map.

We applied the Can-BICS classification to OSM data to create an up-to-date and extensive map of bicycle facilities across Canada using nomenclature that reflects safety and comfort priorities. We completed ground-truthing of the data using aerial and ground-level imagery at over 2,168 locations in 15 test cities.

 Version 2 of the OSM Can-BICS national network dataset data is available on ArcGIS Online. 

Descriptions from Winters et al. 2020 with additional OSM classification details.

We evaluated data quality in 15 test cities (five large cities, with population > 500,000; five mid-sized cities with population > 50,000 and ≤ 500,000; and five small cities with population ≤ 50,000). We considered all street and trail segments mapped as bicycle facilities in open data or OSM. We collected a stratified random sample of 2,168 points. We interpreted ground-level imagery from  Google Street View , aerial imagery from  Google Maps , and other corroborating data where necessary (e.g. PDF maps from the city and newspaper articles). When we compared OSM to what was on the ground, the estimated accuracy for presence or absence was 76 ± 3%, 71 ± 4% for comfort class (i.e., high, medium, and low comfort), and 69 ± 4% for infrastructure type. The confidence intervals for accuracy assessment overlapped between OSM Can-BICS and Can-BICS classified from open data.

This dataset is in the development phase. Accurate classifications depend on both the classification algorithms and the underlying data quality. We encourage interested people to become active OSM editors.  LearnOSM  is a great resource.

Bike paths are under-represented since there is no way to explicitly tag them on OSM. They are often misclassified as cycle tracks or multi-use paths. They compose a very small proportion of total facilities (4% by length). We use a geometric operation to improve the distinction.

Local street bikeways are under-represented. Distinguishing local street bikeways with meaningful traffic calming and diversion from signed routes that offer little to support cyclist safety is a challenge. We continue to develop and refine geometric operations to identify traffic calming and diversionary elements on local streets.

Painted bike lanes are slightly over-represented, since some mapped painted bike lanes do not meet Can-BICS standards in ground-level imagery (for example, a highway shoulder covered with gravel or without bicycle stencil pavement markings or ‘bicycle route’ signage).

The non-conforming category is slightly under-represented. Better differentiation of the above categories will help.

1. Include surface tags to distinguish multi-use paths from multi-use trails. This is critical for accessibility, including for people with electric pedal-assisted bikes and wheelchairs.
2. Include tags for separation between pedestrians and bikes (segregated:yes). This makes the difference between a medium-comfort multi-use path and a high-comfort bike path or cycle track.
3. Map meaningful traffic calming features on local street bikeways using traffic_calming:yes. Consider adding speed limits, and the number of lanes to help people make informed choices about where to ride. These features make the difference between signed routes or shared lanes that only provide nominal levels of comfort and safety compared to high comfort routes that can be comfortable and safe for people of all ages and abilities.

We are continuing to improve OSM Can-BICS. We classified data across the country and are developing metrics for access to bicycle facilities which we will post to  CANUE . Please  use OSM Can-BICS  in your project, and  get involved with OSM !