We are now offering a download of pre rendered tiles up to zoom level 17 on https://tile.osm.ch/tiles.html. The tiles are split up into the cantons as the whole of Switzerland is quite a large amount of Data.
You can download them to either host them yourself, or use them in an offline application.
There was no easy way forward. A little background: Last year, we bough some shiny new SSDs to speed up our servers. I decided to install the new Ubuntu 16.04 onto the new disks, the new long term support version. In the mean time, most services are moved to the server with the new Ubuntu and the SSDs, except routing. I couldn’t get our old version of OSRM to run on Ubuntu 16.04, so I was forced to upgrade to the newest version. But a lot was changed in OSRM in the mean time: the API for routing requests as well as the syntax for the routing profiles. That is the main reason why we still run such an old version. The old user interface we used is no longer maintained, therefore incompatible with the current API. But there is a brand new one, which does not support multiple routing profiles. So starting at the hack weekend in Karlsruhe, I hacked the profile selector into the new interface.
The plan is now to switch to the new OSRM and user interface on Sunday, March 19th.
As you may remember early this year we got permission to use the open address data from the Canton Berne in OpenStreetMap, this after a couple of years back and forth due to the new cantonal legislation and older usage terms that sounded as if the data might be usable but in reality didn’t allow it.
The address data is available from the cantonal open data portal however it is rather unwieldly (a good 400’000 addresses) and problematic to handle even in JOSM. I foolhardly promised to do something about that at the time, but didn’t get around to doing anything up to now. It should be noted that nobody has stepped forward and volunteered to organize an import of any kind, so my focus is simply proving it as reference data that can be used at a small-scale.
While my current solution is not perfect and will likely see improvements over time (for example the layer is currently opaque), it is probably the best solution for now. I’ve produced a background layer from the data that shows
To reduce clutter I’ve shortened some of the usage strings:
Wohnhaus – WH
Bauernhaus – BH
Garage – G
Scheune – S
Gebäude – B
The data is hosted on sourcepolesQGIS Cloud system, many thanks to Marco and his team for supporting us. To make things simpler for iD users and work around some issues JOSM has with WMS servers we actually proxy this through our mapproxy instance.
Take part of this MapRoulette Challenge which is about checking if there’s a pedestrian street crossing (crosswalk) or not. This verification goes on until it’s all done. MapRoulette is a kind of gamified website and a mapping campaign similar to MapZen’s targeted editing series.
So help OpenStreetMap to improve this missing information!
Figure 1: Cycle map from OpenStreetMap main page showing pedestrian street crossings (crosswalks) as yellow dot markers.
A crosswalk node needs to be placed as part of a way. It requires the tag highway=crossing (search for “Crossing” e.g. in built-in online editor “iD”). For the capture of other attributes like the existence of islands or traffic signals see Key:crossing on OSM Wiki.
Important notes: When using editor “iD” the presets suggest tag crossing=zebra. This tag is mainly used in UK. Since this challenge is in Switzerland you can delete this tag. Finally: When saving, a changeset comment is needed. You can use there e.g. “Missing Crosswalks Challenge”.
Figure 2: A “Street Crossing” node shown in the online editor “iD” built-in OpenStreetMap main page.
The street crossings (crosswalks) in fact are coordinates our automated process found in aerial images by using a parallel image detection algorithm. This new software works mainly with yellow crossings of Switzerland on aerial images with certain quality at zoom level 18.
The coordinate data around eastern Switzerland has been extracted at December, 2nd. 2015. This covers about a tenth of the total area of Switzerland.
Contact us if you have questions or ideas around this hot topic.
This article was originally published on geometalab.
On Friday we drove 870 km to Haarlem in the vicinity of Amsterdam to pick up three servers. Wikimedia has decommissioned their two year old toolservers and donated the hardware to SOSM. Thanks to Wikimedia Germany and Wikimedia Switzerland for making this possible.
This will enable us to run our services on our own hardware in Switzerland. Support the move by contributing to the SOSM donation drive 2015.