Skip to content

Instantly share code, notes, and snippets.

@bkowshik
Forked from planemad/validation-osm.md
Created July 13, 2017 07:14
Show Gist options
  • Save bkowshik/b04994f741fed552c053dd6de8960dac to your computer and use it in GitHub Desktop.
Save bkowshik/b04994f741fed552c053dd6de8960dac to your computer and use it in GitHub Desktop.

In the past, we see the future

Welcome to the proposed vision of the OSM data validation logic.

Every new object change is compared to its previous version to understand the nature of change and flag anything that is not normal. An object which breaks the fundamental laws of a world is not normal.

The new and old version of the map object is compared and quantified on various factors:

  • Nature of object (object type, feature type, significance, revisions)
  • Change type (object added, modified, deleted)
  • Tag changes (tags added, updated, deleted)
  • Geometry changes (nodes added, moved, deleted)

The results are recorded as extra properties of the object that can be used for filtering later.

Object type

An object is of only one type during its life, either of:

Point can be anywhere. Usually does not move by more than 500m

Line is an ordered set of points, which when connected will not cross over itself. Usually does not move by more than 100m.

Area is a closed line enclosing only one surface. Usually does not move by more than 50m.

Feature type

  • City has atleast a suburb, or neighbourhood, or tertiary+ road in a 5km radius. Usually does not change names or shift location by more than 1km.

  • Motorway is connected to a road at both ends. One end is always connected to another motorway.

  • Religious place does not change religions or denominations.

Significant tags

  • Name is usually not removed unless incorrect. Names in new languages are added all the time, names rarely change more than once. Names are usually in Capital Case and do not contain expletives or links.

  • Wikipedia is a valid article title on the Wikipedia project

  • Wikidata is a valid data item on the Wikidata project

Importance score

Higher values denotes the object is more important.

  • Tag count denotes how well defined a feature is
  • Tag score cumulative score of weighted tag scores
  • Revisions the number of times the object has changed since it was created.
  • Size the area on the map the object represents
  • Significant location flag denoting the object is located in a geography of interest
  • Connectivity number of other objects physically connected to the target object
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment