[Talk-asia] Grab’s GlobalLogic OSM team edits in Thailand

Mishari Muqbil mishari at mishari.net
Tue Jan 15 16:29:18 UTC 2019


Hi Erwin,

Thanks for going through these edits. I did however find issues with some
of the items that you had marked
<https://gist.github.com/govvin/6c8c9df8106c667b6c28823131419d1e> as Good.

Changeset 64219447 <https://openstreetmap.org/changeset/64219447> way 641456255
is highly unlikely to be correct. The tree in the middle is probably hiding
a wall.
way  641456258 I don't think connects to way 528417189 north of it. It also
retags 528417189 as residential when the original service was more likely
correct.

Changeset 64219422 <https://openstreetmap.org/changeset/64219422> way 641456102
the tree looks suspect, I wouldn't draw this.

Changeset 64218992 way 641450395 tagged as residential when it's a service
going through a temple. Way 641450398 seems to take one through a track (or
a way in disrepair), complete with overgrown grass, but tagged as
residential.

Changeset 64217305 <https://openstreetmap.org/changeset/64217305> way 641425121
is connecting through a wall.

Changeset 63748679 <https://openstreetmap.org/changeset/63748679> way
636777048 is most likely not what it seems

Changeset 63219933 <https://openstreetmap.org/changeset/63219933> way 630873131
is suspect, it goes through a wall

Which means out of those, only changeset I would agree is "Good" is 64219476
<https://openstreetmap.org/changeset/64219476>

This is all I have time for and I didn't look at the ones you had labeled
OK or Bad.

Best Regards
Mishari



On Tue, Jan 15, 2019 at 3:11 PM Erwin Olario <govvin at gmail.com> wrote:

> The original document (in original formatting) from which the following
> text was copied is at: https://hackmd.io/s/HkZmHZ5z4
>
> The following summary is a synthesis of the individual assessments made by
> the volunteer mappers. I am responsible for any errors you may find in this
> document. The invidual assessments are available through their respective
> links.
>
> We are sharing these findings with the rest of the Asian OSM communities
> with the hope of promoting open discussions, and constructive engagement
> for all parties. We feel that many other (Asian) communities can  benefit
> from the exchange.
>
> Kind regards,
> Erwin
>
> P.S.
> Link to Supaplex's review notes are unavailable as of this writing.
>
> ---
>
> Grab operates in Southeast Asia, and their OpenStreetMap editing
> operations are contracted to Global Logic. They have conducted editing
> in several other countries in south-east Asia prior to their foray in
> Thailand. These never came un-noticed by locals and didnt’t always have
> a great start [0], but in time, developed into positive working
> relationship
> with these communities.
>
> In December 2018, an article was published in TechCrunch [1] that described
> the edits made by Grab’s contractor, Global Logic, as “absolute
> disaster.” This tickled the interest of some mappers from several Asian
> communities, and they agreed to conduct their own assessments of said
> edits in Thailand.
>
>
> Changeset assessments by volunteers
>
> The following volunteers agreed to conduct their own indpendent
> assessments of the changesets made by acknowledged team members [2] of Grab
> Logic. These volunteers have extensive experience in OpenStreetMap, and
> are known to be active OSM advocates within their respective
> communities.
>
> Reviewer OSM username Assessments
> -------------------- -------------- -------------
> ABROKWA, Kelvin Muzirian [4] Notes [5]
> CHEN, DennisChen Supaplex[6] Notes [7]
> DWIJANANTO, Adityo Adityo[8] Notes [9]
> OLARIO, Erwin GOwin[10] Notes[11]
> SAMBALE, Maning Maning[12] Notes[13]
>
> The instruction to volunteers was to randomly select changesets using an
> OsmCha filter [3] of un-reviewed changeseets of Grab’s Global Logic team in
> Thailand, numbering about 7,000 edits. And to use OsmCha to verify said
> changesets, and leave comments on them to record observations, if
> necessary.
>
>
> Conclusion
>
> As with any other countries with active mapping volunteers, the Thai
> mapping community is similarly lucky to have their very own active,
> diligent, and prolific mappers who make time to map their favorite
> neighbourhoods. We’ve frequently seen usernames which became familar
> after several changeset validation, and we kept seeing the same names
> again and again. Kudos to these contributors.
>
> Overall, there is no substantive evidence to support the allegation of
> massive detrimental edits, or systematic errors by Grab’s Global Logic
> OpenStreetMap editing team. The errors found appear to be isolated, and
> though there were unusual edits (that aren’t necessarily bad), there are
> some bad edits that should have been caught by their internal Quality
> Assurance (QA) team - the onus of primary validation of their team edits
> should not be on the OSM community, but Grab’s Global Logic editing
> team.
>
> A few changesets were marked as bad, but none were considered critical.
> Those marked as bad should be immediately addressed by Grab’s OSM team.
>
> Some unusual behaviors were noticed:
> * deleting road segments and
> re-adding them;
> * redundantly adding oneway=yes to roundabouts;
> * their JOSM settings appear some users use a different/edited source
> strings
> for the same imagery, instead of what’s available by default from JOSM.
>
> These are unusual, but nothing major or critical.
>
> There are instances of inconsistencies with imagery interpretation:
> tagging, use of various imagery and adjusting for offsets. The mappers’
> training and on-boarding process can improve in these areas, and all
> other mappers, ideally could have re-orientation to familiarize
> themselves with the issues raised.
>
> The Grab’s OSM team should look into supplementing their imagery source
> with Mapillary/OpenStreetCam imagery collected by the local community.
> They’ve been helpful to the validators in many instances during this
> validation effort. Their QA team should take advantage of those, as
> well.
>
> This is not to say that the complaints from the local community are
> unsound - local knowledge trumps imagery interpretation - feedback from
> the community should always be considered, and investigated closer.
> Remote mapping, validation and on-the-ground validation play an
> important role in the mapping process. Grab can do better with engaging
> the local community, or communicating better whatever assistance or help
> they might need. Although, they should be wary of imposing these needs
> on the community.
>
> Validating edits from organized editing activities is primarily a
> responsibility of the project initiators, and Grab’s Global Logic OSM
> data team (as with any other organized editing effort) should be more
> mindful of edits made by mappers with local knowledge when the imagery
> shows something else (especially from trusted and prolific contributors)
> who have this advantage. If in doubt, communicate with the editors.
>
>
> Recommendations
>
> - The internal Quality Assurance team can do better catching the
> issues we caught, and it may seem like a refresher course would be
> very helpful.
> - Give current mapping team members a briefing about the issues
> identified by the individual assessors, and how to address them.
> - Improve communications and engagement efforts with local mapping
> communities, to avoid misunderstanding.
>
> [0]: https://www.openstreetmap.org/user/harrymahar/diary/43671
> [1]:
> https://techcrunch.com/2018/12/19/grab-maps-osm-thailand-southeast-asia/
> [2]: https://github.com/GRABOSM/Grab-Data/blob/master/Grab%20Data%20Team
> [3]:
> https://osmcha.mapbox.com/filters?aoi=739470d2-4cb5-4c40-93b8-205adc88f264
> [4]: http://hdyc.neis-one.org/?muzirian
> [5]: https://gist.github.com/muzirian/04804d765968b0e5764eb0e1213ef51f
> [6]: http://hdyc.neis-one.org/?Supaplex
> [7]: https://
> [8]: http://hdyc.neis-one.org/?adityo
> [9]:
> https://docs.google.com/document/d/1_rI7YuJ5_g8s6HrNUOqRFVoFDUyYxetW0nO1Fk6Fs7A/edit
> [10]: http://hdyc.neis-one.org/?GOwin
> [11]: https://gist.github.com/govvin/6c8c9df8106c667b6c28823131419d1e
> [12]: http://hdyc.neis-one.org/?maning
> [13]: https://gist.github.com/maning/61d9a1962454159b33ffa046bdeb092c
> --
>
> /Erwin Olario
>
> e: erwin at ngnuity.xyz | v/m: https://t.me/GOwin | s:
> https://mstdn.io/@GOwin
> _______________________________________________
> Talk-Asia mailing list
> Talk-Asia at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-asia
>


-- 
Best regards
Mishari Muqbil

twitter: @mishari blog: www.mishari.net phone: +66818240629
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-asia/attachments/20190115/53dda3f8/attachment-0001.html>


More information about the Talk-Asia mailing list