[Massive edit] Fixing typos and other errors in species=* for trees

[Zkir (Urban Eye)]

Hello everyone,

I’m planning to fix some typo errors in species tag, and, since it’s quite a massive edit, I would like to get your feedback before I proceed.

The background:

It’s not a secret that species names are not particularly consistent, but I’ve found very popular values, which are mainly merely typo errors.

• species=Linden Littleleaf

  • Count: 3952
  • What is wrong: It’s English name
  • Should be: species=Tilia cordata + species:en=Linden Littleleaf

• species=Azardiraecta indica

  • Count: 1092
  • What is wrong: typo error
  • Should be: species=Azadirachta indica

• species=Styphnolobium - japonicum

  • Count: 868
  • What is wrong: formatting. hyphen is not necessary.
  • Should be: species=Styphnolobium japonicum

• species=Tilia × europea

  • Count: 831
  • What is wrong: typo error
  • Should be: species=Tilia × europaea

• species=Ligustrum japonica

  • Count: 721
  • What is wrong: wrong Latin adjective ending
  • Should be: species=Ligustrum japonicum

• species=Prunus kanzan

  • Count: 670
  • What is wrong: Species name omitted
  • Should be: species=Prunus serrulata 'Kanzan'

• species=Melia azederach

  • Count: 668
  • What is wrong: typo error: e instead of a
  • Should be: species=Melia azedarach

• species=Acer saccarinum

  • Count: 521
  • What is wrong: typo error, letter h is missing
  • Should be: species=Acer saccharinum

• species=tree

  • Count: 411 (only natural=tree affected)
  • What is wrong: Just nonsense
  • Should be: tag species is removed

There are some more cases, with count<500, but let’s start with ones described above.

Why this matters:
Tags should be consistent and useful for data consumers (Including myself, I would like to use species tag in the UrbanEye3D for trees).

How I plan to do it:
Osm files with the necessary changes will be created by a python script, and then uploaded back to OSM via JOSM.

When I plan to do it:
The change will be applied on Thursday, May 14.

If anyone has suggestions, concerns, or objections, I’d be happy to hear them before I start.

Best regards,
Zkir

[ivanbranco]

No objections. I made a MapRoulette challenge to fix species tags, but most values could be fixed with a mass fix, as you are proposing.

[ivanbranco]

Another easy mass fix could target species tags with one-word values. If that value appears in this list, then species should either be changed to genus if missing or simply removed otherwise.

For example. we currently have 406 instances of species=Prunus

We also have values in the “genus + sp.” form. The logic would be the same: move the value to genus (without the “sp.” suffix) or delete. For example, we have 531 instances of species=Prunus sp..

We even have 411 trees tagged as… species=tree

[Zkir (Urban Eye)]

@ivanbranco thank you for the quick reply

The map roulette challenge is a good thing indeed, and it may serve us yet, but most obvious offences can be fixed in one go

The one-worders are obvious mistakes, but it may be not that easy to fix them [all] automatically.
The first random example I’ve encountered is:

genus=Prunus
taxon=Prunus cerasus
natural=tree
produce=cherry
species=cerasus
leaf_type=broadleaved
leaf_cycle=deciduous

is it safe to just delete species tag in this case? Seems it should be rather changed to species=Prunus cerasus …

We even have 411 trees tagged as… [species=tree]

I guess I can just delete species tag for those ones

[Map_HeRo (Mikke)]

@ivanbranco

No objections.

Same with me. The tagging of genus/species of trees is a bit messy and everybody doing some corrective work there is highly welcome!

@ivanbranco

Another easy mass fix could target species tags with one-word values.

+1 but one has to be aware that also that one word often is wrong. I already have manually corrected things like species=Eiche or species=Linde and the like.

[ivanbranco]

EDIT: Oooooh, I see now. Cerasus is a genus as well

Then the workflow could be:

• species is a single word, and its value is identical to the genus value. Fix
• species is a single word, and the genus tag is missing. Fix
• species is a single word, but its value differs from the genus value. Do not fix, manual check required

[Lumikeiju]

I support this

---

Also maybe worth deciding on this:

Special characters in biological taxonomy-related tag values

It is established, though debated, that values of biological taxonomy-related tags such as taxon, genus, and species are generally to follow the format standardized externally by the scientific community. How should special characters in these values be represented? As an example, I’ll use the Pacific Sunset Maple, Acer truncatum × Acer platanoides ‘Warrenred’ because it contains both a multiplication sign and a quoted epithet. With the International Code of Nomenclature for Cultivated Plants…

TL;DR: For a Pacific Sunset maple (Q127947609)

@Lumikeiju

Should this be entered into OSM:

• With all special characters included
  Acer truncatum × Acer platanoides ‘Warrenred’
• With a multiplication sign, but with apostrophes instead of left and right quotes
  Acer truncatum × Acer platanoides 'Warrenred'
• Without any special characters
  Acer truncatum x Acer platanoides 'Warrenred'

[Senihtu]

good idea. One remark

@Zkir

species=Linden Littleleaf .

What about changing the species-tag but adding a species:en=*. The english name isn’t wrong. Simply overwriting is deleting valid information!

[Mateusz_Konieczny]

@Zkir

species=Linden Littleleaf

• Count: 3952
• What is wrong: It’s English name
• Should be: species=Tilia cordata

that may be valid, but is not a typo fixing

can you change the thread title?
I also suggest to move it species:en rather than remove

@Zkir

species=tree

• Count: 411
• What is wrong: Just nonsense
• Should be: tag species is removed

also not a typo fixing at all
would it be removed only from natural=tree objects?

[ivanbranco]

@Mateusz_Konieczny

would it be removed only from natural=tree objects?

The total instances of that tag are 418. 411 of those are natural=tree, the number Zkir listed.

[Zkir (Urban Eye)]

@Mateusz_Konieczny

would it be removed only from natural=tree objects?

Sure, only natural=tree objects are affected.

I also suggest to move it species:en rather than remove

Ok, if you guys think that species:en=Linden Littleleaf make sense, I will do it that way.

The initial post updated.

[Map_HeRo (Mikke)]

@Zkir

Ok, if you guys think that species:en=Linden Littleleaf make sense, I will do it that way.

To my knowledge the correct british english term for Tilia cordata is “littleleaf lime” whereas “littleleaf linden” is american english.

[SimonPoole]

Given that these values all seem to exist in a larger number, has it been assured that whatever/whoever generated the values has been corrected?

[Mateusz_Konieczny]

assuring this is impossible, for example Node History: 3683985924 | OpenStreetMap is part of species=tree cluster added over decade ago in something what looks like HOT-mapping or similar activity, see overpass turbo

it is impossible to assure it will not happen again, and “HOT mappers do bad imports/add weird unneeded or wrong tags” has not been solved

I guess that I can add them to Dubious tags, MK list | Projects | OpenStreetMap Taginfo but it also does not really assure anything as thousands of dubious one are known, more than enough for people interested in cleanup

making this proposed edit a recurrent bot edit is also an option (I can schedule those, at least for near future)

investigating sources of these tags may be worth doing though, is there some obvious source of mistakes - but it can be also done after cleaning current ones and checking later is any new case appearing

[Lumikeiju]

Hmm, I guess we also should decide on the formatting of the (genus|species|taxon):en=* tags for trees - I have been using US English names in Title Case, ex:

    <tag k="species" v="Tilia cordata"/>
    <tag k="species:en" v="Littleleaf Linden"/>
    <tag k="species:wikidata" v="Q158746"/>

    <tag k="species" v="Fraxinus angustifolia"/>
    <tag k="species:en" v="Narrowleaf Ash"/>
    <tag k="species:wikidata" v="Q518949"/>
    <tag k="taxon" v="Fraxinus angustifolia ‘Flame’"/>
    <tag k="taxon:en" v="Flame Narrowleaf Ash"/>
    <tag k="taxon:wikidata" v="Q128601714"/>

Additional examples

    <tag k="species" v="Acer rubrum"/>
    <tag k="species:en" v="Red Maple"/>
    <tag k="species:wikidata" v="Q161364"/>
    <tag k="species:wikipedia" v="en:Acer rubrum"/>
    <tag k="start_date" v="2006-05-15"/>
    <tag k="taxon" v="Acer rubrum ʽOctober Glory’"/>
    <tag k="taxon:en" v="October Glory Red Maple"/>
    <tag k="taxon:wikidata" v="Q110765852"/>

    <tag k="species" v="Fraxinus americana"/>
    <tag k="species:en" v="White Ash"/>
    <tag k="species:wikidata" v="Q1193369"/>
    <tag k="taxon" v="Fraxinus americana 'Empire'"/>
    <tag k="taxon:en" v="Empire Ash"/>
    <tag k="taxon:wikidata" v="Q115916809"/>

    <tag k="taxon" v="Acer truncatum × Acer platanoides"/>
    <tag k="taxon:en" v="Pacific Sunset Maple"/>

    <tag k="species" v="Acer saccharum"/>
    <tag k="species:en" v="Sugar Maple"/>
    <tag k="species:wikidata" v="Q214733"/>
    <tag k="taxon" v="Acer saccharum subsp. grandidentatum"/>
    <tag k="taxon:en" v="Rocky Mountain Glow Sugar Maple"/>
    <tag k="taxon:wikidata" v="Q15286462"/>

    <tag k="species" v="Acer × freemanii"/>
    <tag k="species:en" v="Freeman Maple"/>
    <tag k="species:wikidata" v="Q9577378"/>
    <tag k="taxon" v="Acer × freemanii 'Armstrong'"/>
    <tag k="taxon:en" v="Armstrong Freeman Maple"/>
    <tag k="taxon:wikidata" v="Q18471079"/>

See: Overpass Turbo - Trees in Seattle, WA, US with (genus|species|taxon):en=* tags

(I am happy to change both the formatting and special characters to whatever is decided by consensus - just let me know )

[Zkir (Urban Eye)]

@Lumikeiju, I am afraid that special characters is not a thing that can be decided. No any standard requires ‘single quotes’ to be curly

regarding multiplication sign there are currently examples of everything:

And it’s not the thing I personally would like to enforce, at least right now.

maybe Platanus xacerifolia is worth changing to Platanus × acerifolia

[Flap_Slimy_Outward]

If you don’t mind, may I add another set that should be replaced?

It’s type=palm on palm tree nodes. type=* has been reserved for relations; thus, they should not be present on nodes or ways. The wiki’s replacement tag is taxon=Arecaceae. Thanks!

[Mateusz_Konieczny]

that should be a separate thread as it is not changing species= at all and should not be assumed to be liked by people who reviewed this thread already

[Zkir (Urban Eye)]

Lets do it as a next iteration, after I commit changes already discussed here (planned for tomorrow).

[ivanbranco]

I didn’t saw this kind of fix in the main post: Node History: 1136100349 | OpenStreetMap

As I wrote, imho it would be better to just move those values to genus.

@ivanbranco

We also have values in the “genus + sp.” form. The logic would be the same: move the value to genus (without the “sp.” suffix) or delete. For example, we have 531 instances of species=Prunus sp..

So in this case:
species=Malus unidentified species → genus=Malus
instead of
species=Malus unidentified species → species=Malus sp.

See also this related discussion where I go more into detail.