Street View coverage in plain English (101)¶
A quick, no-jargon tour of the ideas this tool is built on.
Blue lines vs blue dots¶
Open Street View and you'll see two kinds of blue on the map:
- Blue lines = continuous car coverage. A camera car drove down that road, filming the whole way. These are what tracelines extracts.
- Blue dots = one-off panoramas — usually a photosphere somebody uploaded from their phone at a single spot. tracelines excludes these.
The whole point of tracelines is: keep the lines, drop the dots. We call it the HARD RULE.
Who made the coverage: car vs trekker vs you¶
Official coverage comes from a few sources:
| Who | What | tracelines keeps it? |
|---|---|---|
🚗 Car (launch) |
camera car on roads — the classic blue lines | ✅ yes |
🎒 Trekker (scout) |
a backpack/tripod on paths cars can't reach | optional (--precision drops it) |
| 🏛️ Indoor tripod | inside shops, museums | ❌ no |
📱 You (photos:*) |
a photosphere from a phone | ❌ never |
Honest caveat: Google's data sometimes labels a road-snapped trekker as launch, so "car vs
trekker" isn't perfectly separable. But a phone photosphere is never mistaken for a car line —
that part is airtight. (the rigorous version →)
Why "official-only" matters¶
If you're mapping where you can actually get a street-level view of a road, a random person's single phone photo is noise. You want the roads a camera drove end to end. Filtering to official car coverage gives you clean, continuous lines you can measure, compare, and trust.
What you get out¶
A GeoJSON file — a standard geo-data format — where each road is a line with properties like its source and capture date. Drop it into any GIS tool, or the built-in map GUI. (exact schema →)
Three sources, one map¶
- Google — the densest coverage, but its data has no clean redistribution license (generate it for yourself).
- Mapillary & KartaView — open (CC-BY-SA), community + org driven.
tracelines can pull all three and show where they agree and disagree.
Now go try it → Quickstart: your first extraction