I'm American, but I live in the UK and am currently working on a research project for Network Rail to investigate how machine learning might be used to improve

I'm American, but I live in the UK and am currently working on a research project for Network Rail to investigate how machine learning might be used to improve performance. So I've been reading up on rail networks in general. The Tube is mostly great, but there are issues; the public generally doesn't see them (but I guess you've never been in London at rush hour?). The UK rail network however is a disaster. Privatization has made things worse - organizational chaos because of too many overlapping responsibilities between all the companies / not enough coordination. And the profit motive of course screws up things for what ought to be treated as a public utility. The latest disaster (the re-timetabling of GTR's services) was badly planned, managed, and executed. It's not "government" in general - my opinion is Americans in general are a lot more organized than the British in a fundamental sense. Europe is one a model of exactly how to run things - but at a certain cost to passengers. There is enough buffer space left in the timetable to deal with possible delays or problems, and so services are relatively infrequent, but on the whole punctual. Thus passengers have to adjust their lives for that. The UK however does not do that - buffer space is eaten up to stick freight trains in whenever possible and any issues with them (plus issues with passenger trains) cascade. And UK passengers have always had train services available at the drop of a hat (metaphorically) - the social cost to move over to the EU style of doing things would make a lot of people very unhappy. The GTR disaster is an example of how socially disruptive changes are, especially when badly managed. Japan is of course super organized, so they have train services down to the very second. The UK only labels trains as 'late' if they arrive more than 3 minutes late... As for NYC, yeah, they have a lot of issues. Old tracks, underfunded, but mostly bad management decisions in trying to solve one problem and it snowballs - the NY Times have had some good articles on exactly what is going wrong. It is like no one has been paying attention to all the OR research in this area, but that seems to be normal - the gap between theory and what's actually is done in industry is quite large.