Ask HN: How close together are your bus stops?

85 points by superqwert ↗ HN
In London UK, distances between bus stops are on average between 100 and 200m away from each other.

I'm interested in finding out whether other cities in the world may have comparably short distances between bus stops and whether they have any datasets that could be used to calculate bus stop distance distributions.

105 comments

[ 1.7 ms ] story [ 149 ms ] thread
I guess one could use openstreetmap data here, since they have pretty comprehensive coverage of bus stops, at least in Europe.
Also, national or regional data might be more complete.

If you can speak enough German to navigate the download, you can e.g. find geodata of all bus stops in Zurich at https://maps.zh.ch/?topic=ZVVZH&showtab=ogddownload (they will only deliver to email addresses, so I cannot link the actual data here).

Zürich Tram and Bus distance (beeline) between stops of trips:

Mean 199m

Median 172m

Min 14m (Zürich, Central <-> Zürich, Central Polybahn)

Max 678m (Pfäffikon SZ, Bahnhof <-> Pfäffikon SZ, Schiffstation)

According to a dataset of 1744 trips from https://data.stadt-zuerich.ch/dataset/vbz_fahrplandaten_gtfs

Really cool, and thanks for finding a much more accessible data set than I managed to!

For people non-familiar with Switzerland, though, it needs pointing out that Pfäffikon SZ is no longer in Zurich city nor Zurich state, but Schwyz (thus the SZ). It still comes up in that data set since it is served by ZVV, the Zurich transport authority. It would also be great to know how you group stops together as "the same stop, just different parts of it". I.e. Central vs. Central Polybahn - I would say these are actually colocated, just served by different lines (tram vs. furnicular).

"on average between 100 and 200m" is an imprecise statement. Where does this metric come from?
I haven't got the precise details yet - I've analysed about 20 bus routes so far and they have an average distance between stops of 128.4m. But till I scale up the analysis to all bus routes, I cannot say for certain that it is the London average.
Transport for London has the open tfl API - I'm gathering data from there
200m is incredibly short. I'd say the average in Bangalore is around 700metres. It varies from 500m in dense parts of the city to 1.5km in the not so dense parts.
It's obviously density related, even in London.
It is also obvious that nobody on HN lives in the sticks. Or, if they do, then they drive and therefore have precisely zero interest in bus stops.

My country dwelling parents have a bus stop outside their door. There is one bus per day - weekdays only - and that is only in one direction. This I find mind-boggling. A shopping trip would require an overnight stay in town and for a full circuit of the route to be completed. Somewhere there is a bus driver getting paid for that and a local authority that has mandated this service be put on.

That does seem extreme. That said, a lot of small town and semi-rural public transit seems to operate at a level where it's marginally better than nothing but not much better for those who have alternatives.

Around the exurbs/adjacent small city where I live, there is a small bus service that connects the Walmart, commuter rail, city center, etc. But I'd be pretty sure it's mostly used by those who can't drive.

https://en.wikipedia.org/wiki/General_Transit_Feed_Specifica... is the most used file format. Public transport companies used to submit their information to Google Maps and similar (the 'G' used to stand for 'Google'). You can find those searching for 'GTFS <name of city>' in various open data portals.

Alternatively OpenStreetMap is a huge global dataset. Buses are a small part of https://wiki.openstreetmap.org/wiki/Public_transport It's easy to extract bus stops https://wiki.openstreetmap.org/wiki/Overpass_API (https://overpass-turbo.eu/) but you're probably interested in routes as well. It's a rabbit hole of complexity, e.g. a bus terminal is multiple bus stops, some buses skip stops based on time of day, weekday, season (school buses), some stops are just for drivers to take a break or refuel, direction of route matters.

The transport view on https://www.openstreetmap.org/#map=16/21.3308/-157.8868&laye... is a pretty nice visualization. For specific questions on OSM tools there is https://help.openstreetmap.org/ and a mailing list https://lists.openstreetmap.org/listinfo/talk-transit

I was involved in a project where we could identify Transit deserts (https://en.wikipedia.org/wiki/Transit_desert). Sadly the new administration cancelled this project. We were using GTFS data and US Census data to identify these. Quite interesting stuff. Perhaps the OP is interested in that as well.
London also has bus route 389 - the entire route is I think 1.5 miles, with something like a dozen stops
My rough calculations give it 15 stops for 2420m in direction 1 (avg 160m), and 2235m in direction 2 (avg 148m).

Alas, it appears that route 631 has it beat - 1440m in direction 1 (avg 130m), 1751m in direction 2 (avg 134m)

(although, peculiarly, the information in TfL's bus route download doesn't match what you get from the web site when you look up 631.)

Using TfL open API

-----------------------------

389 inbound:

min: 24.14811016418819

lq: 63.13737031546762

med: 117.45915732519208

avg: 159.1719773263877

uq: 188.0955557392336

max: 845.60445291754

389 outbound:

min: 27.01945592261999

lq: 41.34298814984834

med: 117.00503319935248

avg: 146.97253048021068

uq: 167.28703169991314

max: 809.8962084249061

631 inbound:

min: 7.0860750640801

lq: 27.685480182275676

med: 104.91095883281287

avg: 128.12981380619047

uq: 218.11267557818746

max: 363.41330216008436

631 outbound:

min: 7.636546343123111

lq: 45.34651661429212

med: 105.97332735159804

avg: 135.50969372368004

uq: 203.01187569586028

max: 363.41330216008436

Yeah, I got my numbers from `bus-sequences.csv` downloaded from the API site.
In NYC, it depends heavily on the bus line. Some stop nearly every block (which can be as little as every 100m) whereas some have more widely-spaced stops.

The buses here are notorious for running very slowly (and thus behind schedule), and one of the proposals to speed some some slow buses is to eliminate stops on lines where they are densely clustered. It's become a subject of considerable debate, as residents and in particular older and disabled people are usually against consolidation, whereas most others accept it as a necessary cost of improving service.

The M14, which runs along 14th Street in Manhattan, is an example of where this is happening right now. Currently, the bus spends around 25% of its time standing in stations because it stops pretty much every block. There's a plan to eliminate some of its stops, but it's unclear whether it will succeed. Here's a good read on the subject: https://nyc.streetsblog.org/2019/04/19/never-stop-stopping-r...

I was going to mention NYC too but in terms of the Subway, which is related.

The "local" stops on a lot of Subway lines really shouldn't be there. I'm looking at you 18th Street, 28th Street, 50th Street and so on. 18th Street in particular is 4 blocks away (where 1 block = <1 minute walking) from 14th Street and the entrances to 14th Street will often extend up to 16th Street anyway (with the platform potentially right underneath).

In the other direction 23rd Street is 5 blocks away.

The Subway spends billions of dollars a year on power. Most of that power (for trains) is spent accelerating so every stop counts.

I expect the same people who are against consolidation of bus stops are against consolidation of Subway stops.

Honestly I personally have never taken a bus in NYC despite living here for nearly a decade. Distances in Manhattan are just so relatively short.

50th seems a bit of an anomaly in your list? Otherwise there would be no stop between 42nd and Columbus Circle (I assume you mean the AC line?)
I'm guessing you have no physical issues and don't think twice about walking 10 blocks. I (usually) don't either but that doesn't describe everyone and am able to just grab a cab or a Lyft if not.
There are always going to be competing interests and tradeoffs. Stopping the train is expensive, in terms of time. If one wants to be most accommodating for those with physical issues, then the cost of that is slower or less frequent trains.
Of course. But it's easy for many to unthinkingly look at public transit through the lens of efficiency for able-bodied riders and ignore those who are long-term or temporarily disabled in some manner.
That attitude is part of the reason why few people take public transit in most US cities. When the transit system is not optimized for the majority, the majority will not use it. That leads to a death spiral where it becomes just a safety net system for the people who have no alternatives.
Take the 18th Street stop. You actually can't get in or out of this station without using stairs AFAIK. There are other stations like this. Are you really using these stations if you're unable to walk a few blocks?

It's expensive in time and money to stop and start the trains. Also you need to maintain a station and employ people there.

If you want to provide for those with physical difficulties where walking 4 blocks is an issue, you support those people directly by, say, paying for their Ubers/cabs.

Are you American by any chance? I find Americans are more prone to this particular issue of not being able to make exceptions where necessary. People can't walk 4 blocks? Make everyone wait! Have more stops! Build more stations!

I mean this is exactly why Prop 13 in California was and is ridiculous. If we're really concerned about old people being forced out of their homes by rising property tax bills, why not just give them in particular a subsidized rate? Why do houses in Malibu and Beverly Hills need caps on property taxes because of some retired person in Bakersfield?

> Take the 18th Street stop. You actually can't get in or out of this station without using stairs AFAIK. There are other stations like this. Are you really using these stations if you're unable to walk a few blocks?

I'd imagine for someone in a wheelchair the stairs would present a much greater challenge than an extra four blocks. On the other hand, for someone with a prosthetic leg, climbing a set of stairs while holding a railing is probably far preferable to navigating an extra four blocks of icy sidewalks and streets.

> It's expensive in time and money to stop and start the trains. Also you need to maintain a station and employ people there.

> If you want to provide for those with physical difficulties where walking 4 blocks is an issue, you support those people directly by, say, paying for their Ubers/cabs.

Since both solutions are expensive (and the latter is exposed to exploitation), I'd like to see a cost-benefit analysis (which includes the benefits to non-disabled for having more frequent stops as well).

> People can't walk 4 blocks? Make everyone wait!

It gets paid for in time or money. Maybe it's an American thing to value money more than time? I'm not sure anyone else has to wait, though. They can always walk if it is faster.

This. I’m on crutches right now. I’ll have to take stairs to get in and out of the subway almost anywhere, but that extra four blocks takes a ton out of you.
From a question on the density of subway stops to a rant about Prop 13 in two steps. Impressive. Note to self: research the average number of hops from any topic on HN to Prop 13.
> Why do houses in Malibu and Beverly Hills need caps on property taxes because of some retired person in Bakersfield?

The house my father grew up in in San Jose is now worth multi-millions but my grandmother couldn't possibly have paid the ever increasing property taxes while she was still alive and I think people would have whined about subsidies in well-off neighborhoods like you're doing now. I mean, she lived in that house since the late '40s and Silicon Valley just kind of moved in next door...

That's always the case. Parts of Manhattan used to be absolute squalor, the worst conditions a city had seen, complete slum housing packing 8 people in a bedroom. Now the area has changed and those same 1 bedrooms might go for 4k a month after a power wash and fresh primer. In Detroit, the city used to be very nice and developed instead of derelict before all the rich white people fled to the suburbs. Cities have never not changed, and it's shortsighted to think that they should exist in stasis when that's never been the case in the 10,000 years of history of cities on earth. I find it ironic that the generation which celebrated ripping out entire neighborhoods for interstates and airports and sports stadiums and parking lots complains when someone wants to build an apartment of all things in the lot across the street.

Prop 13 is a bind for people like your grandmother who live on fixed income and who anticipate living at a fixed level. This law insulated them from the housing market, yet they benefit from it's insulation from property tax when they do finally sell their 100k home, in the tax man's eyes at least, for 2m. In that case, surely it would be better to pay your back taxes when you sell, after you've realized your 2000% gain, but that just would make too much sense I guess.

> Are you American by any chance? I find Americans are more prone to this particular issue of not being able to make exceptions where necessary. People can't walk 4 blocks? Make everyone wait! Have more stops! Build more stations!

Americans have no sense of proportionality or optimizing for the system rather than the individual. It makes living in American cities infuriating.

I really like the frequent subway stops. It makes it easy to hop and and off wherever you are. If they were every block like the busses, I would say that's too much, but they're not, they're every 4-5 blocks, and only in Manhattan.

If you took out half the local stops, I'm not sure what the purpose of the express line would be. Of course the express trains would still be marginally faster, but even as is, I was just having a debate with a coworker this morning about whether transferring from local → express trains actually makes sense. (I absolutely think it does, but she made some good points and it's not always clear cut!)

When I visited Shanghai, I noticed that late at night when there are not many passengers, they will kick everybody off the train at a random stop. They do this with a few trains in a row until there are enough people waiting in the station to fill up a train to some acceptable level.

I assume that the point of consolidating everybody into one train like this is so they can avoid wasting power decelerating/accelerating on a bunch of trains. I.e. just run all the empty ones to the terminus without stopping.

It was a bit annoying whenever it happened to us, but I think this is a better system than getting rid of heavily used local stops entirely.

Edit: As a reply explains, it's possible my description is slightly inaccurate.

Would this really save much efficiency versus the much less disruptive option of taking the trains out of commission (for the evening/morning/etc) when they get to the end of their route?
I'm really not an expert on subways, so I don't know. But I can only assume that the people who run the Shanghai metro know what they're doing.
Sorry, shouldn't have framed that as a question directed at you specifically. :)
Same thing happens in Beijing. Where they kick you off happens to be a subway depot where the train will sleep for the night (which isn't always at the terminus, e.g. if it is a loop line or at a remote junction); or it will simply express to the terminus after dumping all of its passengers. Usually, it's just the train's "time" to be off duty, they aren't actually looking at passenger densities or anything like that.
Thanks for the clarification. Good chance that this is what's going on in Shanghai as well. I don't speak Chinese at all so I was pretty clueless.
> They do this with a few trains in a row until there are enough people waiting in the station to fill up a train to some acceptable level

The "acceptable level" probably doesn't account for everyone getting a seat on the continuation train. This used to happen with buses in Wuhan 10 years ago. If two buses on the same route met up and each had only seated passengers, one would send its passengers to the other bus to stand for the rest of the journey. This was annoying because I would travel places at off-peak times just so I could get a seat.

And of course, because this happens occasionally, it affects everyone using the Shanghai Metro because they all need to build in contingency time when planning a trip to get somewhere by a certain time for an appointment.

> The Subway spends billions of dollars a year on power. Most of that power (for trains) is spent accelerating so every stop counts.

The entire MTA spends less than $500 million on electricity annually (which includes the subway, Long Island Railroad and MetroNorth commuter rail) of a roughly $17 billion dollar budget.

Additionally, the 4.5 million people who use the 28th Street stop on the 1 line and the 6.7 million people who use the 28th Street on the 6 line and the 3.3 million people who use the 28th Street stop on the R/W lines politely disagree with you.

Moreover, if power becomes significant you can add regenerative breaking to regain a large fraction of energy (if they're not doing so already; maybe not because of the age of the trains).
The London Underground definitely does that. Also in part because it reduces the heat output. The underground has been heatsoaking for decades and it’s hitting problematic temperature levels.

This article is a great read: https://www.ianvisits.co.uk/blog/2017/06/10/cooling-the-tube...

Apparently about half of the heat generated in the tunnels is/was from braking. And the tunnel temperatures have risen from 14-30C since around 1908.

Is a good read about the engineering challenges.

Annual ridership figures != people. For stops that are primarily commuter stops you have to divide by hundreds to get the actual person figure. 10k people using any given stop for one round trip twice per day is over 7M annual rides for that one stop. It's actually less than that for 18th St; using real data, the average daily ridership for is 8.6k. Compare with a station that is actually busy, like Union Square, which gets 109k daily ridership. Figures from https://graphics.wsj.com/table/NYRIDER_0420

So yeah, those people would be inconvenienced by having to walk a few more blocks, but way, way more people would experience greater convenience by having faster commutes along those very busy lines since they wouldn't need to make such close stops.

There's a happy medium between too many stops and too few. Two stops within a quarter of a mile is way too many. You can imagine a subway system in which a north/south line stopped at literally every block; you'd probably still have more than a thousand daily riders at every stop, but that's no argument to keep every single one of those stops (or even worse, build them). The system performance would be terrible and it'd take forever to get anywhere. The same is still true when you have stops that are only four short blocks away from each other.

I know how the ridership is measured. I’m not sure what your point was. It’s also not surprising that the busiest stations have 10x as many riders as those less busy.

The four track system of the NYC Subway was its most important innovation, and it literally solved the problem your talking about. You don’t want all the stops? Walk to 14th and take the express.

> I know how the ridership is measured. I’m not sure what your point was.

Probably that it's inaccurate for you to say "the 4.5 million people who use the 28th Street stop" because that number is ridership, not people.

C’mon -- that was a rhetorical flourish.
A rhetorical flourish that conveniently made your argument appear much stronger than it actually was.
That literally is the point of rhetoric.
There's an important distinction to be made here -- in Manhattan, in general, there are crosstown buses and buses which go uptown or downtown. And there are 2 kinds of 'block' in Manhattan, 'long blocks' are crosstown (east/west) and 'short blocks' which are north/south. Long blocks are irregular depending on where you are, and are typically much longer than short blocks ~3x.

Most uptown and downtown buses offer limited and local service, with limited making far fewer stops. Most crosstown buses stop at every block.

One more confounding factor is that about decade ago, the city replaced many crosstown buses with fewer but much larger articulated buses without realizing that the behavior of fewer larger buses is to make the average ride slower because of the time spent receiving passengers. To address that they created a new system for paying on certain buses involving buying a ticket before you board. But that is resulting in skyrocketing levels of fare evasion. Unintended consequences.

The M14 is a special case right now because of limited service on the L train (which runs under 14th street) due to necessary repairs. The M14 is getting swamped with excess people who normally would prefer the subway.

Stupid question, probably. But I'm not American.

Is a "block" just a continuous building with no gaps for a road or pavement?

i.e.

   [#########building##########]    [###### building###]  [###]
Would be 3 blocks? Or is a block only delineated by a road, but not a pavement? Or something else?

Can you have "blocks" with a lot of detached houses, I would guess that yes you can? So that makes me think its blocks of possibly many buildings between roads?

In American English, blocks are deliminated by roads and mostly used in dense urban environments, often where the blocks are rectangular.
Etymology note: The word 'blockbuster' was coined by the press to describe the largest conventional bombs used by the RAF in WWII, because they could destroy a city block.
(comment deleted)
A block (which, as you've pointed out is an ephemeral phrase), in NYC at least, is the space along a road from intersection to intersection. It thus usually doesn't count gaps between buildings or alleys or the like. Many of the long (east/west) blocks in Manhattan, for example, have pedestrian pass-throughs in the middle, but that doesn't count as a split in the block.
It's true that the M14 is special right now because it's currently serving riders of the L train, but the proposal to eliminate stops is part of the citywide Bus Action Plan that is being rolled out progressively by borough. Staten Island already redesigned its bus system, and IIRC the Bronx is next.
In São Paulo It is usually between 200m and 600m, ideally 400m.

But it is very irregular. There are lines with up to 1km between stops

I live in Dublin and I take the bus everyday, is mad, I'm talking sometimes less than 50m.

The stop in front of my house and the next one is only three houses away, Georgian style, so no mansions or huge houses.

The transport in Dublin is horrible and is on top 5 of most expensive public transport systems. (No subway)

If you live and work along the Luas it’s not so bad. But otherwise yeah it’s shite.
In Washington, DC, the distance varies considerably, but 200m sounds about right for the average. WMATA does have a "next bus" service that I see people using.
Notably there are also several "MetroExtra" buses that run similar routes as popular bus lines but make about 1/3 as many stops and are therefore much faster.

The NextBus service DC uses is clearly using a very naive algorithm that is often (consistently) too fast or slow. Some of the better apps modify it or add their own data.

Yes, WMATA has managed to move a lot of S9 buses up and down 16th St. NW at busy times of day. Judging from what I see, the S9 is efficient at moving such people as get on, less so in moving numbers of people. It is routine to see three mostly empty S9s go by a stop before a local bus (S1/S2/S4) arrives.
I live in London and I wouldn't say my local bus stops are spaced at 200m intervals.

In Zone 2, sure. Not in the suburbs.

Definitely density related.

In Vienna, Austria stops are (AFAIK) very rarely less than ~300m away from each other, granted I'm horrible at guessing distances so perhaps I'm way off.
Just measured the bus line I used to take from Edgewater, NJ into Manhattan (measured first stop in Edgwater to last stop in Weehawken), and it averages a little over 250m between stops, but they range from around 125m-700m between stops.

One stop that always angered me was my bus stop to the next was 125m. Basically across the intersection, and there was almost always a single person over there, causing us to accelerate then stop.

In Freiburg/Germany (250k inhabitants), the mean distance between bus stops is 634m, more distant than those of the tram (453m). Freiburg has more trams (72) than busses (65) running on 44 km of rails and the sum of all bus routes is 160 km. [1]

Trams run in short intervals and they are convenient when they're near, but switching from tram to bus carries a high time penalty two out of three times.

[1] https://www.vag-freiburg.de/die-vag/ueber-uns

Here is Buenos Aires average is around 200-300m, which I personally find too short. I remember one time seeing two bus stops literally 20-25 meters away from each other. The bus needed to take left on a two lane road so it would circle the block, there was a bus stop just before entering the circle and just before leaving it. Driving distance was ~400m, but walking was a few steps.
For Switzerland, that distance sounds about right, tending more towards 100m (in major cities). There is bound to be a dataset somewhere: being a tiny country means that we map everything in great detail, and the government takes these topics quite seriously. Some links to get you started:

https://www.geo.admin.ch/en/geo-information-switzerland/geod...

https://www.geo.admin.ch/en/geo-information-switzerland/geod...

You can also visualize a lot of the topographic data on map.geo.admin.ch, which is surprisingly useful/flexible, if not very user friendly.

Also, your goberment ask people about stuff like if it OK to deploy 5G in the country ;-)

(I must confess I find that awesome).

Prague

Per [0] (2001; in Czech), average distance between subway stations is 1038m, tram stops 500m and bus stops 698m.

A related metric is walking distance from home to public transportation stop. Per [1] (2011; in Czech), most of people in the city need to walk less than 5 minutes to reach a stop. (There's a table with a breakdown by walking time and city part in chapter 7.) The same work cites average walking speed as 88.69 m/min, which gives you, for most inhabitants, <450m to reach a public transportation stop.

[0] http://envis.praha-mesto.cz/rocenky/DZ_OO/pril_practexty/BK0... [1] https://dspace.cuni.cz/bitstream/handle/20.500.11956/51223/B...

(EDIT: Updated units.)

walking speed as 88.69 m/s

cm/s ?

Czech your math. People there are blazing fast :)
In Warsaw, Poland, the mean distance between bus stops is 1,041 metres. This may seem a lot, but the distribution is heavily skewed by the fact that Warsaw has 4 express bus lines: not taking these into account, the mean is 699 metres.

Source (Polish): https://warszawa.wikia.org/wiki/Autobusy

In Chicago, there are 1,536 route miles and 10,768 stops, suggesting average spacing is about 230 meters. I'm not sure if that double counts stops that serve multiple routes but it sounds about right.
Wow, that's almost exactly what I would have guessed. The block distance in Chicago is 1/8th mile (200m).
In US cities, the densest lines will have one stop per "block" — and that's frequently a regular tenth or eighth of a mile (and more regular as you move towards the newer cities in the west). So you'll typically find minimum distances of 160m or 200m and then multiples of those for express lines or less dense areas.