An OSM map, like the we once looked for after our first contributions and get us addicted!
Hello all. I want to emphasize that we are testing in a sandbox still and will not deploy our edits until we have fully tested all of our changes. As for the question I posted, I believe it was misinterpreted. I was mainly asking if anyone knew of any other validation services external to OSM. I also want to say I have experience in OSM; I am communicating from a separate developer account as per the OSM guidelines, so my history there is not accounting for my existing friends or past changes. We are also working on a vertification process as per requests. Someone also mentioned taking a picture at the bark to get better GPS (I am aware canopies sometimes interfere signals). To solve this issue, we will ask our users to take two photos: one of the full tree and one at the base. The base photo coordinates will be used for mapping into OSM. The users will also have the choice to manually change the GPS point on the imagery basemap if needed. We will also ensure no duplicates are added into OSM. We have the ability to also only allow input from users which have minimum horizontal accuracy. All of our edits will be hash-tagged and labeled as bot=yes. We may also send an email to our users later where they can see their tree and make changes if needed. The images will be of the full tree. Every individual tree attached to each tree node will have a link to the image. Again, we are still figuring out our verification process.
2 Likes
The main concern about data duplication is not related to the user uploading the same tree twice, it is directly related to the user not uploading an existing tree in OSM.
Regarding the correct position of a tree, there are several ways to achieve a minimum precision of location.
However, the project is quite ambitious and you should be ready for a storm of comments as soon as you misplace a tree or mislabel a species.
Have you studied the equivalence of current TAG’s of tree species in OSM against what was found in https://identify.plantnet.org/?
1 Like
I understand your duplication concern. Our method of ensuring no duplicate trees are added into OSM is to find trees that are in OSM that are near any tree we want to upload. If the OSM tree is too close, our data is not added into OSM. As for the species name, we have compared the two. OSM species documentation Key:species - OpenStreetMap Wiki lists common mistakes that we will keep in mind. We are also aware that some species are already common in OSM, such as Platanus x acerifolia, and we will add species in a manner similar to those examples. We are hoping to have a verification team that can ensure the trees are added correctly into OSM to avoid the swarm of comments you mentioned. Of course, if a tree is missed and comments do come, we will try to fix it immediately.
3 Likes
Sorry, another couple of questions for you, particularly in regard to trees being close together.
How are you working Tag:natural=tree_row - OpenStreetMap Wiki , & wooded / forested areas Tag:natural=wood - OpenStreetMap Wiki ?
1 Like
We were not necessarily planning on using these tags. We will request users to only take pictures of trees that are standalone, not part of forests or a group of trees in a row. If a user does happen to send these in, then our verification team can use the two tags you mentioned. The main use, we believe, for our app is for mapping trees in your community/subdivision or near you, not necessarily ones that are part of many trees in a forest or wooded area.
2 Likes
Thanks: this is good, and I had inferred as much. Are you also planning to live test in a limited area before you go “worldwide”?
What I meant was that you should choose just one city, or national park, or whatever, and have a dozen or so authentic contributors (not your developers) doing “live” edits for a while. Those first edits should be reviewed by yourselves, and then your work peer-reviewed by the community, before rolling this out wider. Honestly, it’s the “worldwide” thing that makes me most nervous: it connotes an app unleashed into the wild, and no way of spotting the damage it’s doing before it’s too late. imo the app needs to learn to walk before it can run: OSM doesn’t suit people who move fast and break things. 
Thanks for your clarifications in the rest of that post, too. Especially, I didn’t realise the norm was to use a developer account.
1 Like
That’s pretty much what the iD editor does - if you link to a changeset at osm.org you get a bounding box; link to a node and you get that node as an orange dot.
Given that you know the details of the tree just added you don’t need to go to osm.org to show that Whatever in-app map library you’re using would surely be able to show whatever you like as an overlay.
2 Likes
According to this science paper Mapping tree density at a global scale | Nature there are approximately 3 trillion trees on Earth. 3 trillion of anything on OSM is sure to use a ton of disk space. What would the usecases be for adding all these trees to the map?
I don’t think anyone wants to add “all those” trees to the map. Most trees are part of a forest/wood which are mapped with a landuse/natural area, not individual nodes.
4 Likes
Limited storage space is not a serious limitation to osm editing, at least for any things we map right now.
Mapping trees does not require justification for database use.
(Though this specific import may have other concerns)
2 Likes
Yes we can trial test in a small area and send you sample changesets on here before fullying implementing. We are testing in the master api anyway, so anything we add will not be in the real osm. I have to get back to you with which area we will test in first for the osm community.
Agreed. We do not think our app will even be that popular. Especially since we are only mapping individual trees, not forests, I do not think this will be a problem. The reason I do not think our app will be too popular is because it is only accessible if you have the QR code or direct link, so storage should not be a problem.
2 Likes
Yeah, makes sense. The app would definitely be used more for mapping trees in urban environments; I don’t think that anyone would ever want to map out the individual trees of a forest. Only the nanomappers would do that.
2 Likes
For info, I’ve renamed this thread following a request from the thread starter.
1 Like
I still think that based on my understanding of the app, this is an import. However, if each user of the app were required to have an OSM account and the trees they mapped were added to OSM under their user name, then the app is just another editor and it would not be an import. Perhaps that is the intention and I am just misunderstanding the situation.
1 Like
Seems like this should be an on going Map Roulette project. The organizers only need to create a challenge. Then they ca just focus on collecting and roughly align submitted trees. OSM users can then decide whether to add them as individual objects or woods. This work flow avoids the whole issue of an on-going import.
1 Like
Problem with MR is that people often confirm without actual verification.
MR is good for low risk tasks that can be safely done remotely and require user to take some action manually.
Posting verification tasks is doomed to be blindly clicked “yes” by someone sooner or later.
2 Likes
Are there any other methods that would require more active human interaction? It seems like a better solution than going multiple imports on a semi regular basis.