The same stations and connections, arranged differently. Switch between a hand-tuned 45° schematic and station geography with north at the top. Straight geographic links are not tunnel alignments.
Only the selected metro network is shown. Other transport modes and future extensions are excluded. Station names remain in the local language.
Line or station selection is preserved when switching views. Choosing a station resets the line filter. Geographic labels are available in the list and on hover; they are not all shown at once in the dense centre.
Schematic · All lines
Relative distance change
137.7 pts
Stations: 57 · Connections: 56
For the entire comparison, independent of filters. Higher values mean more changed proportions, not lower quality or a percentage of map error.
Method and limitations
For each connection, measure the straight distance between its endpoints in each view. Divide each set of distances by its own mean. Score = 100 × root mean square of the difference between these normalized distances.
Zero means identical relative adjacent distances; there is no upper limit. Translation, rotation, reflection and uniform scale do not change the score. It ignores angles, bends, label collisions, travel times and readability. Do not rank cities by this number.
Geography uses a local equirectangular approximation: longitude scaled by the cosine of the mean station latitude. This is for compact city networks, not a geodetic calculation. Each line connection has equal weight.
OSM snapshot: 2026-09-13. Topology reviewed: 2026-09-18. Method: relative-edge-chord-rms-v1.