I've seen people use xarray datasets and write some simple loops over it. This works for large datasets as well (xarray can be easily distributed using dask)

Sounds like a fanfic I would enjoy

While this is a great option for personal involvement, I'd like to clarify a few points. Will this approach adequately fund the infrastructure requirements, such as servers, logging, and telemetry services, for a large-scale application? Are there any specific usage limitations?

I share your sentiment. Open web and free web aren't interchangeable. Bluesky's current position as an open and free platform is commendable, but I question its long-term viability without a sustainable revenue model.

My point is that advancing general autonomous driving technology will also benefit public transportation.

... of more advanced sensors and computing power, enhancing their perception of the surroundings and improving their ability to navigate safely. Additionally, public familiarity with bus travel eases public acceptance of autonomous buses and reduces anxiety about the technology.

For the first question, I think that buses are a simpler application of autonomous driving. They are a promising application for autonomous vehicle technology due to their fixed routes, which simplify route planning. Their larger size allows for the integration ...

This is just common sense to me, but apparently it's radical in the US. 🤷‍♂️

Public transport is probably less popular with American investors who never saw how a bus looks from the inside ¯⁠\⁠_⁠(⁠ツ⁠)⁠_⁠/⁠¯

The future of public transport is autonomous. While rail systems are a great start, autonomous buses can revolutionize last-mile connectivity, especially in suburban and rural areas.

Can't a transit system be autonomous? 😄