248. I'm looking at a particularly uniform area in California that has a clear pattern of junctions, corridors, and land use blocks. Before this PR, the classes were a mess:
With this PR, most cases become correct:
Some of the misclassifications will be tricky to remedy. Some more ideas to think about later:
does the cumulative length of roads at different angles inside the bundle tell us anything?
if we look at the connections between bundles, even if they're misclassified, can we start to put together a simplified graph?
248. I'm looking at a particularly uniform area in California that has a clear pattern of junctions, corridors, and land use blocks. Before this PR, the classes were a mess:
Some of the misclassifications will be tricky to remedy. Some more ideas to think about later: