more railcar things
2/5/2025
building on the 30 minute r1+o1 sesh
ok so I played around with a couple more things in a google colab (100% written by gpt4o btw)
- convert the USDOT geojson dataset to a networkx graph
- find nearest graph node to miami and sf
- plot shortest rail route from miami to sf
now the next thing im wondering is: what if we say f a backend to serve this here and we run this in a WASM notebook
also yea if you scroll to the end you'll see what it thinks is the shortest path from miami to sf is by some path that goes through the gulf thats in the USDOT dataset for some reason
am curious what that actually is physically