[Talk-GB] TfL cycle data published - schema mapping

Martin - CycleStreets list-osm-talk-gb at cyclestreets.net
Fri Sep 20 13:02:48 UTC 2019

> https://bikedata.cyclestreets.net/tflcid/conversion/
> I'll shortly e-mail again with more detailed commentary on various 
> aspects of what is shown, in particular cases where new tags are 
> suggested.

As shown on that webpage, most of the data in the CID can be represented by 
OSM tags through a reasonable straight-forward one-to-one mapping. However, 
the following cases are the exceptions.

What are people's thoughts about these suggested new tags?


The schema mapping spreadsheet identifies the following cases where the CID 
has data types that are not ordinarily present in OSM:

• ASL position (left/right/center): OSM represents ASLs, but the entry lane 
has not historically been represented. This may be useful for more advanced 
cycle routing (e.g. a right-entry lane may be unrealistic for a child 
cyclist to negotiate). A new tag, asl_position={left|right|center} is 

• Stepped cycle track: Hybrid cycle lanes have long been the source of 
debate in OSM. The new CID data may provide an impetus for starting to 
resolve this. cycleway:track=hybrid is proposed as a backwards-compatible 
addition that elaborates on cycleway=track.

• Mandatory/Advisory Cycle Lane: OSM has no differentiation between 
mandatory (solid white line) and advisory (dashed white line) lane, 
probably because this distinction is rare elsewhere in the world. A new tag 
cycleway:lane={mandatory|advisory} is proposed as a backwards-compatible 
addition that elaborates on cycleway=lane. This would be useful for routing 
engines, who could infer a level of commitment to cyclists at each such 

• Cycle Lane/Track Priority: This refers to whether the cycle lane 
continues across the sideroad, i.e. has priority. This is the standard 
situation abroad, but sadly in the UK it is not common, resulting in 
arduous and reduced-safety stop-start cycling in the case of a shared-use 
pavement. A new tag cycleway:sideroad_continuity=yes is proposed as a 
backwards-compatible addition that elaborates on cycleway={lane|track}.

• Cycle parking on/off carriageway: This is not represented in OSM, as it 
does not really matter hugely to cyclists. However, this may be useful in 
pedestrian routing, as pavement cycle parking can be an obstruction. A 
simple new tag carriageway=yes is proposed.

• Cyclehoop: No such tag currently (though bicycle_parking=bollard is 
similar), but volume of these in the CID makes this worth trying to 
represent. It represents a form of cycle parking that ideally ought to be 
improved through replacement with real stands rather than retrofitting 
poles essentially to facilitate fly-parking (which can then be disruptive 
to the visually-impaired). Accordingly, advocacy groups may well find this 
useful. A new tag bicycle_parking=cyclehoop is proposed.

• Wheel-rack cycle parking: This is unfortunately increasingly common in 
this UK. It is not currently represented as a specific cycle parking type 
in OSM. A new tag bicycle_parking=upright_stands is proposed.

• Two-tier cycle parking: This is unfortunately becoming more and more 
present in the UK, but bizarrely OSM does not have a representation in 
widespread use currently, with only 6 instances worldwide of 
bicycle_parking=two_tier. It is proposed this be used, which will hopefully 
then galvanise usage beyond London.

• Early-release signals: Currently no support for this, but may be useful 
in improving cycle safety in routing. A new tag 
traffic_signals:bicycle_early_release=yes is proposed.

• Side-road entry treatment: This refers to when a continuous pavement (to 
the benefit of pedestrians) is created across a sideroad, which ostensibly 
doubles as a form of traffic calming, which accounts for its presence in 
the CID. A new tag, continuous=yes, is proposed, to be used in combination 
with sidewalk=yes, as the continuity aspect is really what is important 
here rather than it forming a traffic calming hump as such.

• Sinusoidal shape of traffic hump/cushion: This is a property rather than 
traffic calming type itself. New tag sinusoidal=yes is proposed.


Fields needing significant discussion:

• Cyclists dismount: This is unfortunately a can of worms and may not be 
resolvable in the short term. A Cyclists dismount sign in the UK, which is 
a recommendation that can be ignored (unlike a No cycling sign), does not 
mean the same thing as bicycle=dismount, which refers to cyclists being 
required to dismount and walk their cycle but that a cycle can be legally 
carried (unlike bicycle=no). OSM needs to fix this problem, but it will be 
very hard to do so. (The community also needs to fix the contradictory 
descriptions of bicycle=dismount in the OSM Wiki.) A possible but nasty 
workaround would be e.g. bicycle=dismount_uk representing the UK meaning of 
a Cyclists dismount sign.
https://wiki.openstreetmap.org/wiki/Key:access#Access_tag_values and 

• Two-stage turn signal: This is actually a painted road marking addition 
within the junction, rather than being signals per se. A short spur could 
be added as a new stretch of cycleway drawn in, or it could be presented 
within a relation using the (established) except=bicycle tag, but the 
latter does not then actually represent the presence of the waiting box in 
the adjacent highway coming from the left.


Fields not proposed to become newly represented in OSM:

• Shared Nearside Lane for ASL: This is where a cycle lane or shared bus 
lane becomes a lane for general traffic turning at a junction. Essentially 
this becomes the road, so no special infrastructure is involved and thus no 
tagging is proposed.

• Bicycle parking stand types: The CID has several variations on the 
Sheffield stand, e.g. the M stand. OSM has no such differentiation. 
Introduction of such a variation is unlikely to be entertained, as 
downstream users of the data would not consider the difference important – 
they simply care mostly whether the frame is lockable or not.

• Number of cycle parking stands present: Although the CID captures both 
the number of physical stands and capacity, OSM stores only the capacity. 
There is unlikely to be any practical use for OSM users of the number of 
physical stands, and thus this is proposed to be discarded.

• Physical attributes of signs (colour of any patch around round road 
marking; direction facing): These have no practical bearing on OSM users so 
are proposed to be discarded.

• Sign locations and TSRGD sign number: Around 50% of the database is the 
physical locations of signs, whether they be signs on a pole or road 
markings. OSM is able to represent signage locations, but worldwide this is 
very rare (except for very specific types). OSM is concerned with the 
practical meaning represented by a sign, so the sign pole head is not 
really relevant. Almost always there is another asset, which represents 
this meaning in the direct context of the geometry. E.g. an ‘Except 
cyclists’ sign is an asset next to a cycle lane/track asset whose (line) 
geometry has a contraflow indication. Accordingly, the proposed action here 
is to discard the vast majority of the signage locations which have related 
geometries with meaning-based attributes, and for the rest (e.g. ‘Except 
cyclists’ in a road which has no physical markings/layout but permits 
contraflow cycling, so is not technically an asset and thus not represented 
in the CID) to transfer this meaning onto the road itself.

Martin,                     **  CycleStreets - For Cyclists, By Cyclists
Developer, CycleStreets     **  https://www.cyclestreets.net/

More information about the Talk-GB mailing list