Kernregistratie Topografie Tilburg / Landuse import
Kernregistratie Topografie Tilburg is a partial import of the Kernregistratie Topografie (KRT) dataset which covers (The entire Municipality of Tilburg in The Netherlands). This dataset is also used as a source for the Basisregistratie Grootschalige Topografie (BGT) dataset. The import is currently finished.
Goals
The main goal is:
- Acurately map landuse for the city.
Schedule
Importing is done
Import Data
Background
Provide links to your sources.
Data source site: https://data.overheid.nl/data/dataset/kernregistratie-topografie-tilburg
Data license: https://data.overheid.nl/licenties-voor-hergebruik
Type of license: CC-0
ODbL Compliance verified: yes
Import Type
This import is a one time import. The import is done using JOSM.
Data Preparation
The raw dataset is a single (very large) ShapeFile containing all data of the entire city. So Besides full landuse coverage it also contains the location of trees, streetlamps, benches, drains, etc. The raw dataset is too large to be opened in editing programs like JOSM.
The first step is to filter the raw dataset and export it to smaller files. This filtering and splitting is done automatically using a script made specifically for this import. This script can take the raw data as import, select only the objects needed, en then export the selected objects. It will produce a GeoJSON file in case of complex objects like ways and a CSV file in case of just nodes.
This script also allows for splitting the selected data into seperate areas using a template with areas. For this import a file is used which contains all different areas of the city.
Tagging is also (mostly) taken care of by the same script. Of course it can only tag things if they are in the raw dataset. Missing information needs to be entered manually.
Data Reduction & Simplification
After filtering the data does not have to be simplified.
Tagging Plans
All imported polygons and single nodes are given the tag source="KRT Tilburg". Nodes part of ways are not given a source tag.
Green and vegetation
| Feature | OSM tags | Example |
|---|---|---|
| Grass | landuse=grass | |
| Trees | landuse=forest | |
| Other miscellaneous green
Bits of maintained green that are neither grass or forest. |
natural=scrub | |
| Hedges | barrier=hedge
area=yes |
|
| Fences | barrier=fence | |
| railway | landuse=railway |
Parking
The data has a map of all on street parking areas. Only the areas of parking spaces are mapped.
| Feature | Tags | Example |
|---|---|---|
| Parking area | amenity=parking_space |
Residential gardens
Residential gardens are also imported as landuse. Garden polygons do not overlap buildings or other landuses. They are tagged as follows:
| Feature | Tags | Example |
|---|---|---|
| Garden | leisure=garden
garden:type=residential access=private |
Data Merge Workflow
More important in fitting the data into the existing data in the OSM database. Luckily there is very little existing data. This import generally replaces all existing data. However this has to be checked manually for each case where there is existing data in the database.
Team Approach
I'm doing this solo.
Workflow
Step by step instructions
Import is finished
Changeset size policy
The import will be done in many pieces. Landuse is imported seperately for each landuse type and city area. This keeps the changesets relatively small. Most of the time an upload will fit in a single changeset.
Revert plans
Changesets can be reverted using JOSM. If that is somehow not possible to do it can be done manually. All imported ways and single nodes are tagged with source="KRT Tilburg" so they can be selected using an Overpass Query or filtered in JOSM. Nodes part of ways are not tagged, but are easily detected by JOSM as invalid data.
Conflation
Conflicts with existing data are resolved manually. This will be kept to a minimum as the main goal of this import is to add data where there was none before. If there already exists data of sufficientor better quality in the OSM database it will not get replaced by this import.
The imported landuses will in a lot of cases share borders with existing landuses like buildings. Neighboring polygons will share common nodes where they lie next to eachother. This also includes buildings.
QA
Add your QA plan here.