Uber Pedestrian Data Improvements Around Key Venues (Organised Editing)

General talk
1

Hello OpenStreetMap community,

We would like to inform the community about an upcoming organised editing initiative by Uber focused on improving pedestrian-related mapping around key venues.

The objective of this project is to improve pedestrian map data around locations such as airports, railway stations, transit hubs, shopping centers, stadiums, universities, and other major public venues where accurate pedestrian information is important for navigation.

To ensure transparency, we have published the project details on the OpenStreetMap wiki:

Project Wiki:

The project has also been registered on the Organised Editing Activities page:

Organised Editing Activities:

Countries covered

  • USA
  • UK
  • France
  • Spain
  • Italy
  • Netherlands
  • Sweden
  • Switzerland
  • Germany
  • Belgium
  • Portugal
  • Australia
  • Hong Kong
  • India
  • Brazil
  • Canada
  • Argentina
  • Mexico
  • Greece
  • UAE
  • Saudi Arabia

The project wiki describes the scope, editing methodology, quality assurance process, participating editors, and contact information. All edits will follow OpenStreetMap’s Organised Editing Guidelines and will be made using verifiable sources.

We welcome any questions, suggestions, or feedback from the community. If you have comments or concerns, please reply to this thread or contact us using the information provided on the project wiki.

Thank you for your time and for helping us improve OpenStreetMap.

6 Likes
Unresponsive Uber-affiliated user making incorrect changes
Uber edits in Italia
Uber Pedestrian Data Improvements Around Key Venues - US (Organised Editing)
2

Sorry, but I need to ask: Important for navigating which mode of transport?

Also: Please don’t start right after dumping the little information above here. Give us some time. And make sure the poor clickworkers know that there’s such a thing as changeset discussions and we like the comments there answered and acted upon. Thanks.

6 Likes
3

Thanks for the questions and feedback.

This project is focused only on pedestrian navigation. We’re improving pedestrian-related map data around key venues, such as entrances, exits, walkways, pedestrian paths, crossings, and access points, to help make walking directions more accurate. It is not intended for vehicle routing.

We won’t start the project immediately and will allow some time for the community to review and provide comments.

1 Like
4

Do you plan to start this activities in places where already done Uber edits have unresolved problems?

I see UK being listed, while I know that UK has unresolved problems, see Unresponsive Uber-affiliated user making incorrect changes - #48 by pnorman and the entire thread

12 Likes
5

Also, I would highly advise you to consult local communities. Especially German community. Before making edits there.

14 Likes
6

Please post the exact changes proposed (not a generic “improving pedestrian data around key venues”) on the Canadian forum before executing them. Thank you.

Edit: to clarify, I’m not only requesting this to be difficult. I see on the wiki page that you’re proposing to edit around "Stadiums & Sports Arenas " in Toronto. To my knowledge, these are fairly well mapped already. To be frank, Uber (along with other corporations) does not have a great record of editing. So I would like to know just what you would like to change.

14 Likes
7

Similar for Greece. I see in the project’s wiki the only mention for Greece is for Athens in the section of “Entertainment Districts & Event Venues”.
I may have a suspicion of the places you have in mind, but we need you to clarify the exact places and changes you propose in Greece’s sub-category, because there may be some info that you are not aware of and which may not be obvious from whatever sources you will use (which is why you need to ask each separate community you are about to edit).
As far as I know, many important places with transport hubs in Athens have a pretty good pedestrian-related mapping already.

6 Likes
8

Like the others I see your description as very generic so we do not really get an idea what you want to improve and where.

So I ask you to make a more detailed announcement in every affected country (or area) subforum that is affected in good time before starting your work, so that the local communities are aware of it and have time to react.

For my homecountry Switzerland you mention ZĂĽrich where I feel we already have quite a good coverage in mapping - and also have an active local community so I see a risk that any distance mapping might not really make things better unless the locals are aware of it and can keep an eye on it.

So once again: Please announce your plans, more in detail, in the local parts of this forum before starting.

14 Likes
9

Will you deliver the same quality as Meta has tried?

Given the track record you have in the UK, I hereby ask you not to edit in Germany. The time I and other local mappers would have to spent to review, fix and revert your changes can be used for mapping instead.

13 Likes
10

Uber have made a right mess of their previous pedestrian efforts in London, UK and were unresponsive to changeset comments for a long time.

The changes they made did not comply with their own published plan. The users who made the changes were not the ones listed on their plan.

They have yet to rectify the problems.

I would suggest they dedicate the resources implied in this proposal to rectifying their previous work, then report back.

19 Likes
11

And, only after, possibly come back to local communities before mapping.

5 Likes
12

Which, when you did the same thing last year, & didn’t respond to comments, eventually lead to me reverting 20000 or so of your edits!

As others have said about their countries, please advise us in Australia of your exact proposed changes, 2 weeks before your planned start.

I also wonder about your suggestions, as most Melbourne railway stations are already well mapped, as is the Sydney CBD, which has had a lot of recent construction work done, most of which isn’t visible on any available imagery?

Edit to remove quote that shouldn’t have appeared there? Message was done as a reply to Uber, NOT spiregrain, so no idea why it’s quoting the wrong person?

3 Likes
13

For US activities, I would be a lot more comfortable if this plan made any mention of the Pedestrian Working Group schema and guide:

The guide is tailored to the typical situations in the US, such as recommending that sidewalks are essentially always mapped as separate ways instead of a sidewalk:[side]=yes tag, which is much more usable in jurisdictions like the UK where there is a general right to walk on public streets (except motorways).

3 Likes
16

Thanks for raising this, Graeme.

We’ll make sure changeset discussions are actively monitored and responded to throughout this project.

For Australia, we’ll provide advance notice before starting edits, and we’ll take existing mapping into account. We recognize that many areas, including Melbourne railway stations and parts of the Sydney CBD, are already well mapped, and we won’t make changes unless we can verify that an improvement is needed.

17

Thanks for sharing these resources. We’ll review the Pedestrian Working Group schema and guide and consider them when editing in the US. Where local mapping practices differ, we’ll follow the established community conventions for that region.

18

One thing that irks me with several of these corporate mapping activities is the use of proprietary sources. There’s always talk of “uber gps traces” or “uber street level images”. That’s nothing the community can verify easily and I do not like that.

12 Likes
19

depends on editing quality, Yanosik edits in Poland are based on private GPS traces and they were repeatedly able to update roads before aerial was updated and before any mappers can survey (mapping roundabouts in random places etc)

they probably made mistakes, but error rate was lower than of typical experienced mapper (or mine)

problems start where noticeable part of edits is bogus and you cannot verify them as sources are secret

11 Likes
20

Exactly that’s the point. Uber unfortunately has a record of low quality edits which locals want to double check. So there are a bunch of paid mappers adding data from not accessible sources and local mappers need physical go there and check.

That’s not really a reasonable and healthy situation.

12 Likes
21

The quality of Uber’s pedestrian edits in London was so poor that I would strongly resist any “improvements” they wish to inflict anywhere on the planet.

The London edits were characterised by an absence of adequate training, supervision and QA.

On one occasions, a mapper added sidewalks over a structure which is obviously a bridge and failed to add bridge=yes + layer=* and then unquestioningly accepted iD’s suggestion to create intersection nodes with pedestrian level crossings over the railway lines at the throat of Victoria Station, one of the UK’s busiest stations. A paid mapper who is so poorly trained that they do not know how to map bridges should not be mapping any type of highway, anywhere, ever.

Uber’s mappers also seemed to concentrate their efforts on adding separate sidewalks to relatively minor roads, which can be crossed anywhere and which, without accurately mapping all the driveways etc., tends to have a deleterious effect on pedestrian routing. Moving out into London’s suburbs, we have a lot of quiet residential streets with long Victorian terraces and front gardens too small for a car, so these roads don’t even have driveways to act as routing shortcuts. If a road like this is half a kilometre long, you end up with a pedestrian route to get to the house opposite which is also half a kilometre instead of <10 metres.

Mapping pedestrian infrastructure along busy main roads and through complex junctions with staggered crossings would be useful, but only if it is done by people who know and care what they are doing.

I live in London and use OSM-based pedestrian routing services to create running routes to send to my watch. This year, many of these require adjustment to stop me taking a broken route involving non-existent crossings, or 5 metre vertical movements through concrete bridge decks.

When I fix poor mapping like this, I bitterly resent the fact that I, as an unpaid volunteer, am being diverted from what I actually want to do with my mapping time to fix errors which other people had been paid to make.

Accurate pedestrian information is important for navigation. Checkbox exercises to add decorative sidewalks where they are neither needed nor accurate is an impediment to navigation.

19 Likes
22

and if I guess correctly: iD was not pushing this answer, it was asking whether it is crossing, tunnel or bridge

screen-2026-08-07-08-45-06
screen-2026-08-07-08-45-06379Ă—245 15 KB

so paid editor selected the random (first?) option

3 Likes