0

I want to create an Excel sheet with all residential streets in France. To do so I want to use QGIS and the QuickOSM tool. By doing so I have been running into an issue. QGIS is giving me 4 different data sets when I download residential streets for an area (one with a square, 2 with a line and one with a dot). I'm not sure which one yields the information I am looking for as all 4 datasets have different amounts of observations when I look into their table. Also, when choosing the right dataset, can I be sure that no streets are being double-counted?

7
  • 2
    For this amount of data, you may want to consider using a database. From there you can fetch the road name from the various layers and keep the distinct names by city. OSM data usually comes as nodes (point) or ways (line OR polygon) and some tools try (miserably, in my opinion) to extract roads for you, leading to a 4th layer. This road layer is therefore a subset of the line layer, so the content can be assumed to be duplicated. Check the layer name!
    – JGH
    Apr 2 '20 at 20:14
  • How would I go about using a database in this case? I'm sorry if my questions are a bit basic but I haven't been working with anything but R before so I'm very new to this. Apr 2 '20 at 20:19
  • 1
    One open source option, for which you can find plenty of doc/tutorials, is to use PostGres+PostGIS and the tool osm2pgsql to load the OSM dataset that you can get from geofabrik
    – JGH
    Apr 2 '20 at 20:21
  • And this yields more reliable data then using QGIS and QuickOSM? Apr 2 '20 at 20:26
  • the data is the same. Querying millions of records is much easier in a DB... but that's up to you :-)
    – JGH
    Apr 2 '20 at 20:30
2

For streets you want the line dataset.

The single part lines, usually just called lines.

OpenStreetMap (OSM) has four types of features. Points, Lines, Polygons, and Multi Lines.

Points are used for point objects, but could be part of a street network, like traffic lights.

Lines are used for the streets themselves.

Polygons are used for areas, for example a square, or car park.

And multi lines are used for features that consist of multiple lines. For example a bus route would be captured this way in OSM.

2
  • 2
    Actually OSM has only three different elements: nodes, ways and relations. The types you mentioned are the result of a conversion by a specific software, in this case QGIS.
    – scai
    Apr 3 '20 at 8:04
  • @scai fair point. Apr 3 '20 at 8:53
1

One proposal using Python

I would definitely suggest that you look at OSMnx for downloading street networks if you are familiar with Python. It is a swiss knife, especially suited for dealing with and manipulating road networks.

It is worth a try.

General guidelines

You can for example use the osmnx.core.graph_from_bbox method to use a bounding box for downloading data and appliying a filter that will make a call to the Overpass API for you.

Then you can save your network to an ESRI Shapefile using the osmnx.save_load.save_graph_shapefile method and explore it using QGIS, ArcGIS or any other mapping tool.

It would then be possible to export the attribute table as a *.csv file, e.g. using QGIS:
Exporting attribute table to Excel from QGIS?
https://freegistutorial.com/how-to-export-layer-to-csv-format-on-qgis-3-0/

I let you explore this way upon your wishes.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.