Bus stops here: Shanghai lets riders design their own routes

Even with regular, fixed routes, I've for some time argued the transit operator really need booking apps, on the basis that you really need the data on the full journey, and it'd transform e.g. bus routes if you could offer "there'll be a pickup within X minutes", without necessarily having the buses for it by falling back on renting cars. If you make people give their end destination, you can also do much like what the article suggests, but semi-automatic based on where those on the bus (and waiting at stops) are actually going right now.

Today, ridership gives hard data on where people will go and when given the current availability. Offer a guaranteed pickup, and you get much closer to having data on where people actually would want to go, and even more reliably than people voting on a "wouldn't it be nice if" basis.

I don't even know if my local bus company tracks when people get on and off. It'd need facial recognition to track each person getting on, and when that person got back off the bus.

This is usually done with WiFi MAC addresses. I know that London did this for tube journeys but I'm not sure anybody's done it for busses. You can also use smart card IDs if there is an RFID payment system.

The introduction of randomised MACs might have put an end to it.

This is really a bad idea. I absolutely do not want to explain where I am going anytime I get on a bus or train. In Switzerland, most people just get on because they already have some general ticket for the year or month. And even those that don't, you can just enable 'EasyRide' and as long as that if active, at the end of the day (or when you disable it) it will calculate whatever you used.

And you don't need 'there'll be a pickup within X minutes' because regular bus stops in a developed country already tell you all the buses that will come when. Some like 'Line 1, 2 min', 'Line 9, 5min' and so on.

And for your end to end journey, you can simply open the app and look up your whole journey when you are planning it. If you really don't want to wait a few minutes, you can get there on time.

> but semi-automatic based on where those on the bus (and waiting at stops) are actually going right now.

That's a solved problem with 'request stop'. If its in a city, 99% of the time you stop anyway. For less populated routes, the bus driver can just stop if somebody request its. Its an incredibly simple system that has worked for 100+ years. In Switzerland we even do this for rural trains and it works just fine.

The data companies actually need is this, what bus routes are often full and when. And based on that they can increase frequency.

For example in my city, the main bus line is already really large buses (120+ people) that run every 10ish minutes. And during peak times they run a few extra to increase frequency to 5ish minutes.

In a city, you can run 15min frequency even on the routes that go into the rural area, and for anything else you can do more then every 15min. That fast enough that additional on demand pickup doesn't make much sense.

The most important point is, don't ask people for data just because you want data. If people want to use the app to look up end-to-end journey or buy tickets, that's something you can use. But I sure as shit don't want to open an app anytime I get into a bus, tram or train.

> This is really a bad idea. I absolutely do not want to explain where I am going anytime I get on a bus or train.

So don't. But I want to have the ability to enter where I'm going and get the benefits of better service it could bring. I'm in London - I just tap in with a contactless card, but I'd very happily open an app and pick a destination if it meant I was guaranteed a timely pickup, especially for less well served routes.

I'm all for still letting people get on without indicating a journey; you'd just lose out on the benefits.

> And you don't need 'there'll be a pickup within X minutes' because regular bus stops in a developed country already tell you all the buses that will come when. Some like 'Line 1, 2 min', 'Line 9, 5min' and so on.

I do need that, because buses are regularly delayed, over full and skipping stops. Knowing what the current estimate is doesn't solve the problem.

This has been my experience in at least a dozen countries over the years. You can solve that with over-capacity, but it's incredibly expensive to do so and so won't happen most places. Being able to fix that problem at a fraction of the cost has clear benefits.

> And for your end to end journey, you can simply open the app and look up your whole journey when you are planning it. If you really don't want to wait a few minutes, you can get there on time.

I could. But my experience would be vastly better, if, when I've already looked up the journey, and pressed "go", like I often do with Citymapper for an unfamiliar route, I had a maximum wait for each of those routes.

Not least because if you do this, you could run routes with more dynamic schedule based on demand, and account for unexpected spikes.

> That's a solved problem with 'request stop'.

No, it is not. That tells you when to stop as long as you follow the regular route. If you have information on who is going where, you can dynamically change the routes.

E.g. a route near where I worked often had a very overcrowded leg between two stations. It'd often have served more passengers better to turn some of the buses around at either of those two stations. If you had better data on who were going where and how many people were waiting at other stations, that decision could be taken dynamically, and cars brought in to "mop up" to prevent any passengers from being stranded.

Requesting a stop does nothing like that.

> In a city, you can run 15min frequency even on the routes that go into the rural area, and for anything else you can do more then every 15min. That fast enough that additional on demand pickup doesn't make much sense.

15 minutes frequency is shit. It's slow enough it will cause people to make alternate plans. The routes I would want this on had 8-10 minute pickups and we still regularly ordered ubers for journeys we could do on the bus. The problem isn't when the bus is on time - if I was guaranteed the bus would always show up exactly on time, and never be full, 15 minues would be somewhat tolerable, but the problem is when a delay happens, and the bus that finally arrives is too full to take on passengers.

> The most important point is, don't ask people for data just because you want data.

If you think it is "just because I want data" you didn't get the point.

> I'd very happily open an app and pick a destination if it meant I was guaranteed a timely pickup, especially for less well served routes.

There is nothing about an app that can give you that guarantee. If the system cannot run their current schedule on time data on who wants to go where won't help them. They need to fix their operations to run on time. If their buses are full they need more buses, if they are skipping stops it is obvious that more people want to ride than there is room for without data on who that person is.

Your transit operator already has all the data they need. You need to ask why they are not acting on that data. I don't know if it is incompetence (that would be my expected answer in the US), or they lack the money to run more service. However either way the data they need exists and more data won't help.

Now if the transit operator is competent and has money: more data can help inform what is the best change of all options - but there are better ways to get that data than an app. An app is always limited to those who choose to install and use it (these days phones shut off installed apps that are not in use so you don't get data)

"I absolutely do not want to explain where I am going anytime I get on a bus or train"

And why should the bus driver care about this? You can get off the bus if it doesn't suit you.

Insane idea. You can either tell me your social security number or you can stop commenting on Hacker News.

(Who am I? Well why should I care to tell you that?)

It's not an invasion of your privacy for a bus driver to ask where you're heading.

There is a reason well funded transit agencies do not go into that direct, in fact, the are investing many billions going into the opposite direction. Are all these people just really stupid? That is not to say dynamic information about traffic can't help with things, but its very much an optimization, not the core of the system.

The system you advocate sound really good in your head, because an unknown non-existing system that magically sends a car to any place in 5min anywhere and transports millions of people reliability just sounds fantastic.

You don't see how complex this system would be and how instantly hard this would be to implement, and even if somebody did, it would be more expensive and less efficient, and provide less service and less capacity.

Additionally, most of the problem you complain about, already have known good solutions that could be implemented at far lower cost. And those problems are 100x easier to solve then the new system you are proposing.

And what I really don't understand, is why do you think an bad public transit agency that is already bad at running simple buses, is going to do much better if they had to run a, much more complex highly dynamic system. That is just a contradiction. Just do a simple thing correctly first, follow best practices, and then you can experiment some more with experimental stuff.

> E.g. a route near where I worked often had a very overcrowded leg between two stations

So this is already statically known then ... and all the needed data already exists.

> 15 minutes frequency is shit.

That frequency reliable and coordinated with anything else is for low population areas. If you believe in those areas, a public transport agencies would have cars just ready to pick everybody up, you are fooling yourself. That is just the kind of magical thinking you are talking about.

And 15 minutes is perfectly fine for quite a lot of places, many things in Switzerland run at 15min intervals, and its plenty for many things as long as its coordinated with everything else.

> The routes I would want this on had 8-10 minute pickups and we still regularly ordered ubers for journeys we could do on the bus.

You are likely quite wealthy, because most normal people do not order uber if there is a bus in 8min, even if they are sometimes late or a bit to full for your taste. Because if everybody did what you suggested, uber would be oversubscribed and massive surge pricing would happen and most people wouldn't get an uber, and then only after quite some time.

Since you seem to live in London I would just point out that Britain has done pretty badly on transit. For mostly dumb, tourist, reasons they are sticking with double decker buses. These are exactly the wrong solution for most routes. Slow ingress and egress in double deckers increases dwell times, and that's a killer as it leads to bunching and station skipping.

London is far to big a city for these tiny buses on all but a few routes into the outskirts. What they should use modern trams or something like this:

https://www.bus-pics.com/pics/_data/i/upload/2020/06/23/2020...

This can transport 120+ people and much more if you really pack in people.

And lines where this is overkill or not possible because of other constraints, level boarding electric/hybrid bus with many doors and a single level are the right alternative.

These are just some of the many issues with how London runs its system that leads to some of those problems you describe. I'm not an expert on London, but I'm sure people have written about this.

I have lived in Zürich and Berlin, and I have only once in my live skipped a bus or tram because it was to full (because of a fire at the train station). And Zürich has a 96-98% on time rating for buses and trams, and even higher if you account for how often you make your connection, I have only once in my live take an uber in the city, and that was at 3am. And Zürich is still considered a quite car oriented city, and doesn't have a metro like London, where buses and trams often run in traffic and and there are far to fewer bus lanes then there should be. Even some roads that have 4 lanes and run many buses still allows car on all 4 lanes for some dumb reason. But you can plan a reliable network even with that.

For a good general article going over many of these topics, this is a nice one and has some good further links:

https://marcochitti.substack.com/p/getting-bus-priority-righ...