Latvia/Riga address point import
- 1 Goals
- 2 Schedule
- 3 Import Data
- 4 Data Preparation
- 5 Data Merge Workflow
- 6 QA
Improving addressing in Riga city, capital of Latvia, by importing address point for missing addresses. Address point dataset is provided by Riga municipality via open data website. Address point import will significantly improve address information in Riga, especially in poorly surveyed parts of city.
The import should be launched in one changeset in end of May 2015.
- Data source site: https://opendata.riga.lv/ (in Latvian)
- Data license: https://opendata.riga.lv/licence.html (in Latvian)
- Type of license: ODC-By
Prepared and reduced dataset includes around 22 thousand nodes of address points.
OSM Data Files
Link to prepared import .osm data is not ready yet.
One-time import performed by importing single .osm file via JOSM.
Data Reduction & Simplification
Main data reduction work is performed by excluding building addresses and address points which are already in the OpenStreetMap database from the dataset, provided by the municipality. Due to changes in OSM data since import data was prepared, the import data will be updated prior importing.
Provided dataset contains only corresponding to "addr:street" and "addr:housenumber" tags. For all nodes the "addr:city=Rīga", "addr:country=LV" and "source=opendata.riga.lv" tags will be added.
- source: Rīga municipality open data website opendata.riga.lv
- comment: Import of municipality provided address points in Rīga.
Assumed data transformations are minimal and include:
- transformation of LKS-92 coordinates to WGS84 coordinates of nodes
- renaming street and house number tags to appropriate
- addition of addr:city, addr:country and source tags
- modification of street names, which have mistakes in municipality dataset (very few)
Data Transformation Results
Link to OSM XML files will be prepared.
Data Merge Workflow
The import data is prepared with help, recommendations and advices of Latvia OSM community, and is assumed to be imported in a single changeset by one dedicated OSM user.
General import workflow includes:
- Preparation of reduced dataset (is done)
- Discussion in Latvia OSM community (is done, advices and support is given)
- Fixing mistakes in the existing OpenStreetMap Riga addresses (typos in street names) (current stage)
- Import data update prior to import
- Data import via one changeset
Prepared changeset once was imported into OSM (but later reverted) thus changeset size is not a limiting factor.
In case of errors and mistakes during import, data imported with mistakes is assumed to be reverted fully.
All addresses in Riga city, which are already in OSM, are excluded out of final import dataset. Sporadic conflation mistakes can occur due to mistakes in OSM existing addresses, but from first (failed) data import these mistakes are individual, don't affect overall addressing quality improvement, and will be solved manually later.
Pre-discussion (reverted) import of dataset showed that everything went fine, and no large mistakes were found, except some small conflation errors, which should be minimised prior to official import.