Removing wrong trees / forests / named nodes in Quetzaltenango, Guatemala?

Hi,

I stumbled upon a lot of names on the map of Quetzalenango, Guatemala. It turns out that the city has a lot of nodes with “natural=tree” or “natural=wood” which also have names. In a changeset comment I saw that this might be some attempt at organized mapping gone wrong.
In addition to wrong trees / woods, there are also nodes who only have a name tagged, which also fits the pattern. Overpass turbo shows the extent of all those nodes.

I checked the aerial imagery available as well as mapillary for some of those streets. It is definitely bad mapping, there are not this many trees along the roads. It looks as if the waypoints of gpx tracks have been uploaded and turned into “trees” or “woods”.

The mappers who were participating seem to have been active only for a very short time, none have responded to previous changeset comments (all links to HDYC)
AdanH
Ana Zuñiga
Ixquik Lucia
Linda_28
marvin colocho
Roger Abel
Saira16
Saulopm
Zelma Sosa

Before I started writing here, I thought I would just ask if I should delete the nodes.
Now I think a complete revert of all the contributions of those users might be better. There have been some attempts to “correct” the natural=woods into natural=trees, but apart from that, I hope there are not too many conflicts. For a complete revert of all their changesets, maybe someone from the Data Working Group has some tools to speed up the process?
(Ping @woodpeck ?)

What do the mappers in LATAM think, just delete all nodes like that, or make a full revert of all the contributions from the accounts above?

Sorry for not writing in Spanish or Portuguese, I hope the translation tool is working fine :slight_smile:

1 Like

OK, I just saw that @marczoutendijk already reverted most or all changesets from user Saira16, see the discussion on this changeset.
Some nodes must have survived, since Overpass found them.

1 Like

Hello,
I just could try a DWG revert for all the edits of the users in your first mail, including Saira16.
Do you want me to perform such a revert?

Marc Zoutendijk
OpenStreetMap Foundation
Data Working Group

1 Like

Hi Marc, I think that would be best. Maybe wait a day to see if someone from the local community wants to chime in.

2 Likes

Ok, I’ll wait another day and shall perform the revert on october 7, 2022.

Marc Zoutendijk
OpenStreetMap Foundation
Data Working Group

1 Like

Please do the revert, as a user I already tried to contact users on some changeset but didn’t get a response from them, already asked how long do I have to wait to request a revert but got different responses so if you help us with that. I will really appreciate it.

1 Like

I am also a local editor from Guatemala. A while back ago, I too tried contacting some of the people involved in this tree quest - and even offered to help to add them using the correct tags. I never got any response, so I doubt there would be an issue in doing this revert. Thank you!

2 Likes

As promised I reverted all changesets mentioned in post 1 (@daganzdaanda ) in two reverts:

Because a bunch of trees was changed from natural=wood to natural=tree by:
LSkalayo | OpenStreetMap (for #Kaart 2022)
we now have been left with 1412 trees with the name=XXXX tag that couldn’t be reverted with the help of the reverts scripts.
You can see them all by running the Overpass Turbo script in post 1. (see my next post)
I can simply remove them all with another script, but I’m waiting for reactions from the community.

Marc Zoutendijk
OpenStreetMap Foundation
Data Working Group

1 Like

I used this Overpass to find all trees that were changed by LSkalayo:

[out:json];
(
node(user:"LSkalayo")[natural=tree][name]({{bbox}});
);
out meta;
>;
out skel qt;

The script in post 1 includes a few nodes that shouldn’t be deleted.

Marc Zoutendijk
OpenStreetMap Foundation
Data Working Group

1 Like

I had noticed these contributions while reviewing inconsistent tagging of ways and nodes as natural=tree and wood, and I commented on a small sample of these changesets (e.g.). Before me LSkalayo from Kaart had already intervened, and the result is roads with very many trees. maybe someone else wants to check.

The names assigned to these natural=wood nodes are in the form <sequential-ref>, <description>, and descriptions are one or more of ‘cultivo’, ‘cesped’, ‘flores’, ‘otro’, or ‘nada’. Some sequential numbers show typing errors, like the sequence 2145, 2146, 247, 2148 … as if they never reviewed their own work.

I would guess it’s flower beds, possibly empty and left as bare ground (nada), or just grass (césped), or possibly guerrilla gardening (cultivo). It’s a pity that we don’t manage to get in contact with the individuals nor the organization behind them. As organization, I would guess some public local institution.

1 Like

See Changeset: 127216042 | OpenStreetMap as concluding action for the reported issues.

Marc Zoutendijk
OpenStreetMap Foundation
Data Working Group

1 Like

Thanks a lot! I’ve cleaned up the last remnants with Changeset: 127223542 | OpenStreetMap
Those mostly were nodes which were connected to some geometry and had nothing but a “name” tagged. I only removed the name and left the geometry alone.

2 Likes

Thank you all for acting on these, I stumbled upon these trees when preparing the Mapillary Update a few months ago.

Great job everyone :slight_smile: