[Talk-sn] Import data from ARD-Saint-Louis, Senegal
augustin.doury at hotosm.org
Dim 30 Juin 16:34:15 UTC 2013
Hi Skora, Pierre and Andrew,
Thank you for your answer!
I reply to Skora here but your points are evoked too (admin_level on a
type:point & boundary=administrative on the relation + members ways)
I've imported "communes"(type point) and "Chef Lieux des Communautés
Rurales" (type point) yesterday and finished the import today with "Limites
des Communautés Rurales".
>From rural communities, I've deducted boundaries for the three departments
of Saint-Louis region (Saint-Louis, Dagana, Podor) and for the 7
arrondissements (Ndiaye, Mbane, Cas-Cas, Gamadji Sarre, Saldé, Rao, Thillé
Boubacar) and I've edited it in OSM. I put a note on the relations linked
to Cas-Cas and Ndiaye arrondissements because I have not found their
admin_centers (Cas-Cas village and Ndiaye village).
Response below :
On Sun, Jun 30, 2013 at 2:15 AM, william skora <skorasaurus at gmail.com>wrote:
> Hi Augustin,
> As someone who has worked with HOT and studied in St. Louis, Senegal, I am
> excited to see this happen. I have a couple questions and suggestions:
> - It's generally preferred to use a create a specific user name for
> imports and not to use your own account.
> a user name like 2013ARD_STL_IMPORTS for example, would be appropriate.
Too late for that, sorry to the community. According to the guideline :
"the user's editing history is a permanent record of the source and doesn't
interfere with tags or increase the size of the database as much".
Import is really small so I guess it's ok but thank you for this
information which is pertinent and good to know.
> As I understand, it's not necessary to add the tag source=ARD_STL;CSE to
> each way if you add the source=ARD_STL;CSE to the changeset.
I added the tag source=ARD_STL;CSE to each way.
Administrative boundaries already present in Saint-Louis Region (borders
with Louga Region and Mauritanie) were always tagged with source=*,
stipulating an URL linked to the original data.
Because different sources are used in the same region (ARD_STL;CSE is the
third one) for admin boundaries, I think it's easier to know the source by
a simple click instead to download history.
Once again, I agree with the fact that this process is heavy for the
database, not really in this case because the import is small. But if you
think that it shouldn't be done on principle, this would be really easy to
> For the *"communes" (type point)" -* As nodes*, *they do not need to have
> admin_level=8 to them. I think just place=* , name=* would be sufficient.
I hesitated for a longtime on this point.
"Commune" is a special admin status for a village or town or city.
> place=* and name=* don't mention this specific status wich is the same
admin_level as "Communautés rurales" (admin_level=8).
I thought that admin_level=8 helped to keep this information and would be
useful when official administrative boundaries would be known.
Moreover, use of admin_level=* on a type:point don't seem to be incoherent
according to TagInfo<http://taginfo.openstreetmap.org/keys/admin_level#overview>
Once again, this would be really easy to delete if necessary.
> Regarding the *
> *shapefile "Limite_Communauté_Rurale"
> *I am not sure if it necessary to add the tag boundary=administrative and
> admin_level=8 to the member ways. I have seen boundaries where only the
> relation has tags (for example
> http://www.openstreetmap.org/browse/way/219920348) but I also have seen
> boundary=administrative tagged on the relation as well as the relation - as
> described in
> *which is listed as a reference in the SN tagging guide -
I think both methods are valid.
You give the good links and reasons to this choice here :
"I also have seen boundary=administrative tagged on the relation as well as
the relation - as described in
*which is listed as a reference in the SN tagging guide -
> * *
> Most importantly,
> If you could send a small example .osm that would consist of the data that
> you wish to upload, I would greatly appreciate it and would be very helpful
> to better understand the import.
I saved some work during edition, it's linked to this email as a .osm file.
Becareful, this data is obsolete. Only 16 of the 18 rural communities were
edited. Departements and arrondissements were not edited.
> These are just my interpretations of OSM community consensus and existing
> practice, I hope to hear others' opinions and interpretations as well.
> Will Skora
Très bonne soirée!
-------------- section suivante --------------
Une pièce jointe HTML a été nettoyée...
Plus d'informations sur la liste de diffusion Talk-sn