Just to make life a little more fun, the parking:lane:*=no* tags have been moved to parking:condition:*=* (proposal approved on 2022-01-21). Unfortunately, this doesn't work properly in Osmose.

osm.wiki/Key:parking:condition

Thanks!

Thanks!

A section of service road parallel to Hertford Road appears to be one way in the wrong direction. Does the way need reversing, or is the cycleway a contraflow?

osm.org/way/669759917

Pedestrians, yes, so it should probably be something like foot=destination. Bikes could physically get round on the pavements, but I don't think I'd want to explicitly tag access for them.

There appear to be two "barriers" on Rewell Street in the TfLCID traffic calming dataset.

RWG043305 is an unspecified barrier (TRF_BARIER=TRUE). After looking at TfL's photographs, it's a gate across the carriageway, but leaving the sidewalks unobstructed. There's no evidence that bicycles are permitted, so perhaps:
barrier=gate + vehicle=private + foot=yes
https://cycleassetimages.data.tfl.gov.uk/RWG043305_1.jpg
https://cycleassetimages.data.tfl.gov.uk/RWG043305_2.jpg

The other feature, RWG043304 is an side road entry treatment (TRF_ENTRY=TRUE). I would probably have mapped it as something like:
traffic_calming=table + highway=crossing + crossing=unmarked + crossing:island=no + tactile_paving=yes + kerb=flush
https://cycleassetimages.data.tfl.gov.uk/RWG043304_1.jpg
https://cycleassetimages.data.tfl.gov.uk/RWG043304_2.jpg

Bing imagery here is not very clear and the gate cannot be seen in available Mapillary images.
https://www.mapillary.com/app/?pKey=297077145411938

Why barrier=yes? TfLCID entries have two photographs showing what sort of barrier is present, if it cannot be determined from database fields. How can a barrier of unknown type have known access restrictions?

Barrier n9748911676 appears to be TfLCID asset RWG065518, which appears to be a rising bollard from tfl's survey image dated 2017-09-07. This was already mapped as n6682086798. Unfortunately, TfL's surveyor has incorrectly placed the location of the traffic calming infra to the East of the junction with St Luke's Close, which is easy to confirm as there's a rather distinctive church in the photograph.

A similar set of bollards is present to the East of the junction of Mitchell Street and Helmet Row. This is not recorded in TfLCID, but clearly visible in one of the photographs for the speed table at the junction mapped as RWG064984.

https://cycleassetimages.data.tfl.gov.uk/RWG065518_2.jpg
https://cycleassetimages.data.tfl.gov.uk/RWG064984_2.jpg

I have changed the section of Mitchell Street between St Luke's Close and Helmet Row to highway=pedestrian, deleted n9748911676 and added the TfLCID refs in osm.org/changeset/123046010

As all TfLCID entries appear have two photographs taken by their surveyors, there is no excuse for not using that information as part of the conflation process. Many location also have Bing aerial imagery and Mapillary images more recent than TfL's surveys, although in this case trees obscure the aerial imagery and there is nothing in Mapillary.

It also seems absurd that the imported data does not include the TfL ref or the check_date.

@AyushS183 there are two sets of Mapillary images on that section of street. The link you posted is to imagery from 2015!

Please could you confirm whether or not your edits are part of the TfLCID conflation done on behalf of TfL by Sweco UK Limited?

@AyushS183 The problem is that the tool you have used to "verify" the import shows that (advisory) cycle lanes *were* present on this segment of Chobham Road in October 2017, when TfL's surveyors checked assets RWG154882 and RWG155596.

Even the most cursory examination of Bing aerial imagery (date uncertain, but certainly more recent than 2017) shows that the footways/sidewalks have been widened and have cycle pictograms indicating that they are shared cycle tracks.

There is also Mapillary imagery from February 2020 available, as pointed out in my original comment.

Mapping historic "low quality" cycle infrastructure like advisory lanes, when it has been replaced with "better" infrastructure with physical separation, is somewhat unhelpful to users wishing to use OSM maps or data to plan cycle routes.

In the light of recent legislative changes regarding the enforcement of parking and moving traffic contraventions in mandatory cycle lanes (cycleway:lane=exclusive), not importing the type of lane from TfLCID may also be a questionable choice.

Checked on the ground 2022-06-26T12:45. Corrected in osm.org/changeset/122865564

The southern end of Water Lane (w1064067306) is tagged as both cycleway:both=no and cycleway:left=lane. Please could you clarify which you believe to be the case?

I see that you have added cycleway:right=share_busway to Lea Bridge Road (w585538279) between Onra Road and Melbourne Road.

This section of road was mapped, by a mapper who has actually been there, as lanes=2 + cycleway=separate. As there are only two lanes for traffic and no PSV/bus lane mapped (because there isn't one any more), the undiscussed import of a historic feature without adequate checks is somewhat unhelpful to the quality and accuracy of the map.

#DWG

The cycle lanes which you added to Chobham Road (w10818748) no longer exist, as examination of Bing aerial imagery and Mapillary clearly show.

The TfLCID data is now several years old and needs careful conflation wit existing OSM data, not an undiscussed import which overwrites mapping by contributors who have actually surveyed the infrastructure and checked the imagery.

https://www.mapillary.com/app/user/rskedgell?lat=51.549139944722&lng=-0.0030227655555556&z=17&pKey=176240357710782&focus=photo

#DWG

Thanks for changing the obsolete cuisine=vegan|vegetarian, but surely these should be replaced with diet:vegan|vegetarian=only in most cases?

That looks fine, thanks for updating it.

Hi, thanks for adding the LTN filter. It's probably worth adding the tag bicycle=yes to the pedestrian segment, assuming the restriction there is "no motor vehicles".

osm.org/way/1070316703

The aquatics centre extends at least part of the way underneath the pedestrian area in order to accommodate the two training pools. Perhaps w163734632 would be better tagged as building:part=yes than building=*

The street relation looked useful when I discovered it on the wiki, but a quick look at taginfo shows only 56 in GB, which may include the 21 I have just deleted.

Just out of interest, what was the query which pulled the street relations as well as the highway ways?

To be honest, I'd forgotten they existed! I think all the QA tools try to match addr:street on a POI with the name of the nearest highway=*. Hopefully they'll catch up with parentStreet etc. (if they haven't already).