3 ms·

They have a concept similar to what you described in South America. In Peru you can jump in private vehicles that run particular named (sometimes even numbered)

by matt3D 3y ago

They have a concept similar to what you described in South America. In Peru you can jump in private vehicles that run particular named (sometimes even numbered) routes. The car and minibuses simply go when they are full or once they have waited a fixed amount of time and they have at least on passenger. It works pretty well and a company called CityMapper tried to do something similar in London. They looked at the transport data and found routes that people wanted to take, but we're underserved by public transport (eg. Shoreditch to Islington on a Friday night). They would then hire privately operated black taxis to run the route for a fixed fare with a fixed waiting time, and they took the risk on empty seats. https://techcrunch.com/2017/09/21/citymapper-ties-with-gett-to-launch-shared-taxi-commuter-route-in-london/ https://techcrunch.com/2017/09/21/citymapper-ties-with-gett-...

vidarh 3y ago
I live in London and use Citymapper all the time. I think the problem with their tests is that to do that well you need bigger scale and a very long term commitment, because e.g. Shoreditch to Islington would serve 1) an existing demand, 2) that are on Citymapper. In London that'll work in the centre, because people will know there will be some route, even if suboptimal (Shoreditch to Islington right now, granted it being a Sunday which makes it worse, is 30 minutes on public transport and 39 minutes walk, it's certainly a stupid gap). But the real social benefit is when you find real dead spots, where people likely aren't searching on CityMapper that often because they are familiar with (live, work) in the area, and know they're in a bad spot for buses and have few alternatives, and so the potential demand goes un-spotted. E.g. most of the buses near me goes in an approximately radial pattern out from 3-4 of the local train stations for quite a while before they start turning in various directions. That leaves some areas where you're better of taking a car or a long walk to the nearest station because you're halfway between two bus routes and almost equidistant from the station, and many routes requires going in to the nearest station and out along a different radial. A handful of routes cutting across those radials, and ideally connecting two stations, but in an arc, would make opting for public transport far more viable for a lot of people living in the area. An experimental route like Citymappers would get some people to use it opportunistically But the real payoff comes when you offer a commitment to keep running that route for long enough that people start relying on it: Maybe they opt against buying that second car (or first), or move there on the basis of good transport links. That takes a willingness to commit to an elevated level of subsidy for something like a decade or more to really get the payoff. And that's also why cutting the cost of introducing a new route could make such a huge difference. In terms of finding areas, instead of just looking at "hot routes", if trying to do this with data from something like Citymapper, I'd also try to correlate a population density map with below average number of searches for that density, and then look at what those people are searching for when they still do (and here we also have the tragedy of siloed data: data from e.g. Uber would be far more likely to reveal underserved routes, because they capture a whole lot of searches from people who already know they're poorly served by public transport for specific routes; e.g. I could get to my sons school by bus, but I know it'd be very slow, so if I need to go to a parents evening etc. I'll take Uber; but getting data out of Uber for this would likely be hard, unless perhaps you'd manage to offer them an exclusive or preferential deal on being the provider of "backup cars" to fulfill waiting time guarantees)

by matt3D 3y ago

vidarh 3y ago

I live in London and use Citymapper all the time. I think the problem with their tests is that to do that well you need bigger scale and a very long term commitment, because e.g. Shoreditch to Islington would serve 1) an existing demand, 2) that are on Citymapper. In London that'll work in the centre, because people will know there will be some route, even if suboptimal (Shoreditch to Islington right now, granted it being a Sunday which makes it worse, is 30 minutes on public transport and 39 minutes walk, it's certainly a stupid gap). But the real social benefit is when you find real dead spots, where people likely aren't searching on CityMapper that often because they are familiar with (live, work) in the area, and know they're in a bad spot for buses and have few alternatives, and so the potential demand goes un-spotted. E.g. most of the buses near me goes in an approximately radial pattern out from 3-4 of the local train stations for quite a while before they start turning in various directions. That leaves some areas where you're better of taking a car or a long walk to the nearest station because you're halfway between two bus routes and almost equidistant from the station, and many routes requires going in to the nearest station and out along a different radial. A handful of routes cutting across those radials, and ideally connecting two stations, but in an arc, would make opting for public transport far more viable for a lot of people living in the area. An experimental route like Citymappers would get some people to use it opportunistically But the real payoff comes when you offer a commitment to keep running that route for long enough that people start relying on it: Maybe they opt against buying that second car (or first), or move there on the basis of good transport links. That takes a willingness to commit to an elevated level of subsidy for something like a decade or more to really get the payoff. And that's also why cutting the cost of introducing a new route could make such a huge difference. In terms of finding areas, instead of just looking at "hot routes", if trying to do this with data from something like Citymapper, I'd also try to correlate a population density map with below average number of searches for that density, and then look at what those people are searching for when they still do (and here we also have the tragedy of siloed data: data from e.g. Uber would be far more likely to reveal underserved routes, because they capture a whole lot of searches from people who already know they're poorly served by public transport for specific routes; e.g. I could get to my sons school by bus, but I know it'd be very slow, so if I need to go to a parents evening etc. I'll take Uber; but getting data out of Uber for this would likely be hard, unless perhaps you'd manage to offer them an exclusive or preferential deal on being the provider of "backup cars" to fulfill waiting time guarantees)