Cognised existence: Transport networks are the physical corridors along which people and goods move. Their geometry and classification determine routing, accessibility analysis, and connectivity modelling.
Question: What transport networks exist at a location and how are they classified?
OSM wiki: https://wiki.openstreetmap.org/wiki/Key:highway | https://wiki.openstreetmap.org/wiki/Key:railway
Realisations
Instead of hardcoding implementation schemas here, SPHERE separates semantic meaning from dataset implementation. See the following realisations for how to access this data:
- [[Realisations/OpenStreetMap_transport_networks|OpenStreetMap —
highway=*,railway=*,aeroway=*]]
Natural Earth — roads, railroads, airports, ports
Natural Earth provides small-to-medium scale transport infrastructure layers covering the entire globe. These are suitable for continental overview mapping and context layers, not for routing or sub-national analysis.
Download: https://www.naturalearthdata.com/downloads/ → Cultural
| Layer | Scale | Geometry | Content |
|---|---|---|---|
ne_10m_roads | 1:10m | LineString | Major roads globally, field type (Major Highway, Secondary Highway, etc.) |
ne_50m_roads | 1:50m | LineString | Generalised road network |
ne_10m_railroads | 1:10m | LineString | Main line rail, field type (Rail, Ferry) |
ne_10m_airports | 1:10m | Point | Major civilian airports, field iata_code, name, scalerank |
ne_10m_ports | 1:10m | Point | Major seaports, field name, website, scalerank |
Python load
import geopandas as gpd
roads = gpd.read_file("ne_10m_roads.zip")
railroads = gpd.read_file("ne_10m_railroads.zip")
airports = gpd.read_file("ne_10m_airports.zip")
ports = gpd.read_file("ne_10m_ports.zip")
# Filter to major airports only (scalerank 0–2 are the largest)
major_airports = airports[airports["scalerank"] <= 2]
major_ports = ports[ports["scalerank"] <= 2]Important: Natural Earth transport layers are context reference data. Do not use them for routing — they carry no topology. For routing, use OSM via osmnx.
Geometry Representations
| Rep ID | Source Dataset | Geometry Type | Native CRS | Suitable For | Not Suitable For |
|---|---|---|---|---|---|
transport_osm_graph | OSM via osmnx | NetworkX graph (LineString edges) | EPSG:4326 | Routing, isochrones, accessibility, network analysis | Simple visual mapping without topology |
transport_osm_lines | OSM via Geofabrik | LineString | EPSG:4326 | Visual mapping, length statistics, buffer analysis | Turn-by-turn routing without topology processing |
transport_ne_10m_roads | Natural Earth 10m | LineString | EPSG:4326 | Continental overview mapping, context layer | Routing, sub-national road analysis |
transport_ne_10m_railroads | Natural Earth 10m | LineString | EPSG:4326 | Continental overview mapping, rail context | Routing, detailed rail network analysis |
transport_ne_10m_airports | Natural Earth 10m | Point | EPSG:4326 | Airport location mapping, index of major airports | Airport capacity, traffic, or operational data |
transport_ne_10m_ports | Natural Earth 10m | Point | EPSG:4326 | Port location mapping, index of major seaports | Port throughput or vessel traffic intensity |
Important: Select transport_osm_graph (osmnx) for any routing or accessibility analysis. Raw line features do not carry topological connectivity.
Limitations
- Speed limits (
maxspeed) are inconsistently tagged in OSM — validate before use in routing models. - Turn restrictions exist in OSM but require explicit processing.
- Natural Earth transport layers cover only major infrastructure and are not suitable for local or sub-national analysis.
- For legally authoritative road classification, prefer national datasets.