Suggestion: new category for trees?

[ivanbranco]

Hi, I noticed there are many trees that have species= and/or species:wikidata=, but don't have leaf_type= and leaf_cycle= tags yet. Often this makes a quest appear on StreetComplete, or such trees are marked as incomplete information on Every Door ecc. wasting contributors time to add an information that is - potentially - already there.

Do you think it could be a good idea to have a new category for trees, that based on species= and/or species:wikidata= values could suggest the correct leaf_type and leaf_cycle?

This could also help finding errors, such as a species=Ginkgo biloba tagged as leaf_type=needleleaved ecc., so it could serve as QA as well.

[LaoshuBaby]

I find this difficult and we may need help from a botanist as there are some common trees that are not easy to identify specific species

[ivanbranco]

not easy to identify specific species

I don't understand. We would use already identified species. If someone tagged a tree as species=Ginkgo biloba, NSI would suggest to add leaft_type=broadleaved and/or leaf_cycle=decidous and/or species:wikidata=Q43284

If someone tagged a species=Ginkgo biloba as leaf_type=needleleaved it would suggest to fix it with leaf_type=broadleaved and so on.

[1ec5]

This is a duplicate of #94.

I see what you mean about the incomplete tagging creating work for StreetComplete/Every Door users. But are we sure that NSI-generated iD presets are the best solution for that issue? For most if not every species, the leaf cycle and leaf type are always the same across all individuals of the species, so an automated mass edit or MapRoulette challenge would be just as accurate and much more efficient.

Ginkgo biloba and palm trees probably deserve their own special presets in id-tagging-schema just because they’re so unusual. That is, if anyone can agree on how to tag palm trees…

[1ec5]

there are some common trees that are not easy to identify specific species

This is probably a reference to common names like, say, “pine” or “banana” that can refer to various species in a genus or family.

[bhousel]

I don't understand. We would use already identified species. If someone tagged a tree as species=Ginkgo biloba, NSI would suggest to add leaft_type=broadleaved and/or leaf_cycle=decidous and/or species:wikidata=Q43284

This is a good point, and something we could do with NSI right now, if people would find it useful.

Here are the current common species values in OSM: https://taginfo.openstreetmap.org/keys/species#values

Under our current practice of collecting values used more than 50 times, we'd collect about 1500 species from the OSM planet file, then volunteers would need to match these to their appropriate species:wikidata and other tags, or filter away the ones that are tagged wrongly. (Maybe add locationSets too, for where these things grow? or leave them all as worldwide, because they can live in arboretums or indoors? I really don't know.)

Personally, scrolling through this list, I don't know what most of this stuff is, so I can't judge whether the tag values are useful or not. For me, it's not like brands or transit operators where I can look up what these values are. But I'm definitely open to the idea of adding a "species" tree to NSI if other people want to do the work.

[ivanbranco]

so an automated mass edit or MapRoulette challenge would be just as accurate and much more efficient.

Isn't this true also for brands? An automated mass edit could add brand:wikidata to every McDonald's without it. The problem is that automated mass edits have guidelines to be followed, and MapRoulette challenges of this kind are considered automated edits as well.

Here are the current common species values in OSM: https://taginfo.openstreetmap.org/keys/species#values

Let's take the most used value, species=Acer platanoides. Just adding this NSI entry would affect 60091 elements! Only 678 elements have both species:wikidata, leaf_type and leaf_cycle, it's the 1.2% of the total. And this is 1 species value alone. Imagine the improving this check could have on the database in trees mapping.

[LaoshuBaby]

Under our current practice of collecting values used more than 50 times, we'd collect about 1500 species from the OSM planet file

Maybe 50 occurrences is too low a threshold for a tree, but even with 1000 occurrences, it's 10 pages (it seems that standardizing 200-300 plants in the early stage of the project is an acceptable amount of work)?

[ivanbranco]

I have 50 entries ready (more will follow in the next days according to my free time) to start populating the database in case the suggestion will be implemented:

species;species:wikidata;leaf_cycle;leaf_type Acer campestre;Q158785;deciduous;broadleaved Acer ginnala;Q15634290;deciduous;broadleaved Acer negundo;Q161166;deciduous;broadleaved Acer platanoides;Q26745;deciduous;broadleaved Acer pseudoplatanus;Q156944;deciduous;broadleaved Acer rubrum;Q161364;deciduous;broadleaved Acer saccharinum;Q158301;deciduous;broadleaved Acer saccharum;Q214733;deciduous;broadleaved Acer tataricum;Q162728;deciduous;broadleaved Aesculus hippocastanum;Q26899;deciduous;broadleaved Alnus glutinosa;Q156904;deciduous;broadleaved Betula pendula;Q156895;deciduous;broadleaved Carpinus betulus;Q158776;deciduous;broadleaved Castanea sativa;Q22699;deciduous;broadleaved Celtis australis;Q255375;deciduous;broadleaved Cocos nucifera;Q13187;evergreen;broadleaved Corylus colurna;Q148939;deciduous;broadleaved Cupressus sempervirens;Q147513;evergreen;needleleaved Eucalyptus leucoxylon;Q161468;evergreen;broadleaved Fagus sylvatica;Q146149;deciduous;broadleaved Fraxinus americana;Q1193369;deciduous;broadleaved Fraxinus excelsior;Q156907;deciduous;broadleaved Fraxinus pennsylvanica;Q161164;deciduous;broadleaved Ginkgo biloba;Q43284;deciduous;broadleaved Gleditsia triacanthos;Q157650;deciduous;broadleaved Juglans regia;Q46871;deciduous;broadleaved Liquidambar styraciflua;Q469652;deciduous;broadleaved Olea europaea;Q37083;evergreen;broadleaved Picea abies;Q145992;evergreen;needleleaved Picea glauca;Q128116;evergreen;needleleaved Picea pungens;Q146025;evergreen;needleleaved Pinus mugo;Q147475;evergreen;needleleaved Pinus sylvestris;Q133128;evergreen;needleleaved Platanus orientalis;Q161105;deciduous;broadleaved Populus nigra;Q147064;deciduous;broadleaved Prunus avium;Q165137;deciduous;broadleaved Prunus cerasifera;Q146951;deciduous;broadleaved Prunus serrulata;Q165321;deciduous;broadleaved Pyrus calleryana;Q3079266;deciduous;broadleaved Quercus ilex;Q218155;evergreen;broadleaved Quercus robur;Q165145;deciduous;broadleaved Quercus rubra;Q147525;deciduous;broadleaved Robinia pseudoacacia;Q157417;deciduous;broadleaved Sorbus aria;Q157960;deciduous;broadleaved Syringa reticulata;Q1683340;deciduous;broadleaved Taxus baccata;Q179729;evergreen;needleleaved Tilia cordata;Q158746;deciduous;broadleaved Tilia platyphyllos;Q156831;deciduous;broadleaved Tilia tomentosa;Q161382;deciduous;broadleaved Tilia x europaea;Q163760;deciduous;broadleaved

edit:

8 more:

Acer palmatum;Q269224;deciduous;broadleaved Abies alba;Q146992;evergreen;needleleaved Abies balsamea;Q428023;evergreen;needleleaved Abies concolor;Q145939;evergreen;needleleaved Abies nordmanniana;Q148920;evergreen;needleleaved Aesculus × carnea;Q163779;deciduous;broadleaved Ailanthus altissima;Q159570;deciduous;broadleaved Albizia julibrissin;Q750307;deciduous;broadleaved

[1ec5]

Isn't this true also for brands? An automated mass edit could add brand:wikidata to every McDonald's without it.

This isn’t true for all brands. Just look at all the NSI entries that require locationSets or disambiguators in their names. Anyways, I’m not arguing against pairing species with species:Wikidata. However, I think leaf_cycle and leaf_type are rather poor justifications for opening the door to trees in NSI, when Wikidata already has properties for collecting this information with more nuance than OSM tagging could possibly allow. These keys are just a crutch for 3D renderers that can’t be bothered to look up QIDs in Wikidata or an ad hoc lookup table.

Let's take the most used value, species=Acer platanoides. Just adding this NSI entry would affect 60091 elements!

Would it be only one NSI entry? What would be the preset’s name (since this is the name suggestion index)? If the preset is simply named Acer platanoides, no one but a botanist would find it. If we name it “Norway maple”, then only English speakers would find it, while Spanish speakers in Spain would see English all over the preset list.