Clearance: Quality filter for OpenStreetMap replication

Clearance is a free software tool for controlling the quality of OpenStreetMap replication diffs. It tracks thematic and territorial edits to OSM and keeps replication extracts (extracts, diffs, and a local Overpass API) up to date.

Instead of trusting every incoming change, Clearance evaluates edits against configurable quality rules based on OSM tags, metadata, geometry and changeset properties. Compliant changes, at object level, pass through automatically. Suspect ones are retained rather than applied. Retained data must either be fixed directly in OSM or approved manually. All data contributions are made only in OSM itself. Reviewing and fixing suspect changes is done collaboratively by the team responsible for a given theme and region.

Because it uses standard OpenStreetMap ecosystem formats for both input and output, Clearance integrates seamlessly into existing OSM data reusers workflows, while providing greater confidence in the consumed data. It is used to filter and review changes on thematic contributions such as tourism POIs across France, or road and emergency access points in Spain.

How Clearance addresses this problem

Clearance imports an initial OSM PBF extract, then checks every incoming change against configurable quality rules. Changes that meet the rules are applied automatically to the replicated database, while suspect changes are held back. Quarantined changes must then be fixed directly in OSM or manually approved by reviewers. After each update, previously held objects are re-evaluated, so a change is released automatically once it no longer fails the rules.

… See full entry

We (Teritorio.fr) publish an approach and an implementation of a strategy to reconstruct OpenStreetMap object history at a semantic et geospatial level, beyond the technical object history.

It should help to review the OpenStreetMap changes. It supports geometrical and also semantical OSM object spiting or semantic concept moved from one objet to an other.

A demo is available online https://teritorio.github.io/openstreetmap-logical-history-component/

And the the code is open source : https://github.com/teritorio/openstreetmap-logical-history

The full story on

• Github https://github.com/teritorio/openstreetmap-logical-history/blob/master/doc/OSM-Logical-History-From-technical-diffs-to-reconstructed-semantic-object-history.md
• the same, on Medium https://medium.com/@frederic.rodrigo/osm-logical-history-from-technical-diffs-to-reconstructed-semantic-object-history-eea089a6d72e

Experiment on Disputed and Claimed OSM boundaries

I made an attempted to display boundaries according to some countries official point of view. Most of boundary are consensual. But some are de facto with no recognition from one or two of the neighbors. This facto are the ground mapped in OSM. But some countries have different political claims over the territories. Display this claim is often mandatory in these countries. So some boundaries or territories are disputed by countries. Disputed here meaning, that the country is not officially recognized the de facto ground status. Some other boundaries or territories are claimed. The idea here is partially based on the rejected proposed features “Mapping disputed boundaries”.

Consensual borders does not require more details.

Disputed boundaries are here treated as way level. As only a part of the de facto boundary is rejected.

Claim over territories are treated on boundary=claim relation as extent of the facto boundary relation. boundary=claim relation are not closed relation. De facto boundaries relation, minus disputed way, plus boundary=claim should done a closed relation.

In this experiment I do no try to rebuild this closed polygon. I based on osmborder to derive a modified version. Originally, the tool de-duplicate the border from the relations to ease the render.

So in the output of the tool, a way of boundary is marked as: * neutral, if is part of the standard de facto ground boundary framework, the standard OSM meaning of boundaries. * disputed, if has disputed=yes, dispute=yes, border_status=dispute, disputed_by or boundary=disputed tag, plus if is part of a boundary=claim relation. Typically can be rendered with dash. Note, a neutral de facto can be disputed. * disputed_by, list of countries code rejecting this way as a boundary. * claimed_by, list of countries code claiming this this way as boundary.

… See full entry

The biggest step in this period is the coverage of new countries:

• Full Africa support
• Europe: a number of new small countries: Azores, Albania, Bosnia and Herzegovina, Bulgaria, Croatia, Cyprus, Estonia, Hungary, Isle of Man, Kosovo, Latvia, Liechtenstein, Lithuania, Macedonia, Malta, Montenegro, Moldova, Romania, Serbia and Slovenia.
• Europe: a new analyser server in the Netherlands, country divided in regions.
• Southeast Asia is a new server in Germany covering Thailand, Brunei, Burma, Cambodia, Laos, Malaysia, Singapore and Vietnam.
• Add countries in Asia: Sri Lanka, Syria, Tajikistan, Turkmenistan.
• Add Central American countries: Belize, Cuba.

Countries that are not supported by new servers were supported on OpenStreetMap-France or Iceland ones. Now there is a Coverage Map on Osmose.

Meanwhile, new translations came Portuguese, Polish, Japanese and Hungarian.

On the back side engine, fewer bugs and more unit testing and improving mechanisms for testing. Also, added support for a variable number of objects involved in an error.

For data integration proposal analysers, redesign the code format to go to something more descriptive and structured, so less code, more configuration and convention. Improved data integration proposal analysers to add a “dynamic” support. It means manage a large number of different tags for the same OpenData set (typical French sports fields that have 180 different combinations of tags for the same dataset).

On the side of analysers them self, not so many news. This period was a time of maturation, improve the quality and “conquer the world”.

… See full entry

Cette demi-journée était le lancement public du processus d'ouverture des données en région Aquitiane. L'initiative rassemble à la fois la région le département de la Gironde et la communauté urbaine de Bordeaux. Cet empilement semble orienté vers la métropole locale au détriment du reste de la région. Mais l'approche est inverse. Il s'agit d'une initiative parti de la communauté urbaine s'étendant aux autres partenaires.

Parmi les invités l'on a retrouvé des acteurs nationaux et locaux pour faire un état des lieux.

On a peu entendre l'APIE parler du futur portail national. Ouvert aux différentes institutions y compris locales pour diffuser leurs données. Néanmoins la commercialisation de ces données et l'idée que des données diffusées gratuitement peuvent être réutilisées sans licence en vertu de la loi est soutenu. L'intervention laissait entendre la possibilité d'une redevance pour des usages non commerciaux et même dépassant les coups de mise à disposition pour des usages commerciaux.
Hors intervention, je me suis laisser dire que l'APIE étudiait le cas de OpenStreetMap et la condition de non dénaturation des données spécifique à la loi française. La conclusion serait que de part ça nature de wiki, et donc grasse à ça traçabilité, le projet puisse utiliser les données sans accords spécifiques sur la dénaturation de la part des fournisseurs.

L'on a aussi pu voir la représentante du projet sirtaqui du comité régional de tourisme aquitain. Ils veulent ouvrir les données mais avec des conditions et limitations digne d'un dédale. Relevant plus du marchandage que d'un esprit volontariste d'OpenData. Par exemple avec des niveaux de licences différentes pour les données multimédia et de la négociation aux cas par cas par l'échange de « bons procédés ». Arguant que l'ouverture seule n'allait pas être suffisante à la promotion de leur données, imposant ainsi des limitations aux réutilisateurs....

J’arrête là les critiques et je passe aux réjouissances.

… See full entry

Bonjour,
Pour faire suite aux récentes discutions sur la mailing-list talk-fr sur le projet d'association, ou pas et de la possibilité d'intégrer l'OSGeo-fr, un sondage à été mis en place.

http://doodle.com/3nm3y5czf3ukdgtg

Compte rendu des discutions sur le sujet sur le wiki :
osm.wiki/WikiProject_France/Projet_d%27association_en_France

Information sur l'OSGeo-fr :
http://www.osgeo.org
http://wiki.osgeo.org/wiki/Francophone

De l'avis d'une majorité, la communauté OSM française est parvenu au stade ou sa constitution en association loi de 1901 devient souhaitable. Plusieurs possibilités s'offrent à nous :
1) créer l'association de toutes pièces,
2) rejoindre l'association OSGeo francophone qui poursuit des buts similaires
(géodonnées et logiciels libres),
3) attendre,
4) ne rien faire,
5) nsp,
6) autre.

Vous aussi donnez votre avis : http://doodle.com/3nm3y5czf3ukdgtg

Français :
J'ai enfin fini par écrire l'article sur OSM sur le site de adella.org. À l'origine je mappais un peu uniquement pour écrire cet article, puis je me suis pris au jeu. L'article est fait, et moi je suis encore là ;-).

English:
Finally I wrote the small paper about OSM on adella.org site. I went here initialy to do it, but I do lot of map stuff before write it.

http://www.adella.org/spip/spip.php?article83 (fr)

Count way, node and relation from bordeaux area (only last element editor counted) :
138 kalon33
165 ceyquem
228 Welshie
243 gqjtcbit
275 zorgzorg2
306 moore33
316 evrard
549 FredB
697 Xav
910 andrewpmk
3532 frodrigo
3701 yota