On high capacity systems, train dwell time becomes the limiting factor on passenger capacity. 30 seconds is generally the minimum possible dwell time a system can manage, 20 seconds might be possible during much lower demand periods. But you’re unlikely to ever achieve better than that.
The London Victoria Line, which runs with 90 second headways at peak, achieves at best 24 second dwell times in central section, but 30-40 seconds is more realistic for most stations.
Don’t forget, dwell time includes more than just passengers getting off and on. It has to include time to open the doors, close the doors (including a 2-3 second visual and audible warning!), perform needed safety checks, and eventually pull away. Those operational components the sandwich the core “people getting on and off” bit of station stops add up to a non-trivial number of seconds.
As a minor note, the NYC subway uses a system of fix automatic blocks, not a moving block system, which is what your simulation appears to use. Moving block systems are dramatically more efficient than fixed block systems. But I have no idea how you would get hold of accurate block locations for the NYC subway.
I want to love this but the visual language makes it kind of unusable for me. Why not match the track and train colors to their line color (red for the 123) and then use different visual indicators for train state (stopped, at station, etc)?
For example:
Selected: Black fill
Normal operation: Color fill with 100% opacity
Slowing down: 70% opacity
Stopped: striped fill, 50% opacity
At station: pulsing opacity
Sometimes when I Edit Routes and click an "<- Add" button I get the console error "Uncaught Error: coordinates must be an array of two or more positions" and the page blanks out.
I just tested it out. Since this uses OpenStreetMap data, it is very easy to adapt to other cities.
All input data and scripts are in src/data. Run the two Overpass queries, replacing New York with a city of your choice, and you will get JSON data on the infrastructure, as well as services. Replace the two existing JSON files, run the two transformation scripts, and you are done.
That is actually the idea. In this simulator, you design services that run on existing tracks. NYC subway is unique in the world in that there are a ton of interlining as well as quadruple tracks, so there are plenty of opportunity to improve system efficiency without building anything new.
I can't figure out how to add a route from Bay Ridge to Clinton in NJ. That's my wishlist subway - got to Newark Airport from Brooklyn without going through Manhattan.
Edit: Actually Clifton is in NY, and playing with Google maps there's ZERO public transport from Staten Island to NJ, except by going through Manhattan!
So my idea wouldn't help anyway, unless they extended that subway all the way over the Goethals Bridge.
As a Brooklynite I wholeheartedly agree. Unfortunately this is more of a service designer as opposed to a track designer, and it only works with existing infrastructure :(
There are 4 bridges: Goethals, Outerbridge, Bayonne, and Verrazano. You can walk on Goethals and Bayonne. I was actually thinking a few weeks ago about walking from Elizabeth to Bayonne someday, just for the heck of it.
Looks like a 2 or 3 mile walk from the nearest public transport stops that I can find. Maybe someone local can find a closer stop.
Forest Ave/Goethals Rd North to Bayway Ave at Mckinley St.
Edit: This could be a fun game - find the two spots with the greatest ratio of driving vs public transport. I think those two are pretty good candidates.
Distance: 2.5 mile.
Drive 9AM: 5-9 minutes, 11PM: 4-6 minutes.
Public transport 9AM: 1:52, 11PM: 3:34.
Ratio (using the average drive time) 9AM: 16, 11PM: 42.8.
Another game: when is public transport slower than walking?
According to Google Maps it's not in this case - I'm getting a walking time of 14:37 (!), where the "walk" includes a ferry from Weehawken to Manhattan and the Staten Island Ferry. However Google Maps doesn't seem to know you can walk across the Goethals bridge.
It's not the bridge, the starting point seems to be on a freeway or something, if you move the starting point slightly toward the bridge it will show proper walking directions.
This is sadly not mobile friendly in an amusing way; the map controls are nice and legible but they block basically the entire map. If there was a way to minimize them (perhaps automatically), this might work well on mobile.
On high capacity systems, train dwell time becomes the limiting factor on passenger capacity. 30 seconds is generally the minimum possible dwell time a system can manage, 20 seconds might be possible during much lower demand periods. But you’re unlikely to ever achieve better than that.
The London Victoria Line, which runs with 90 second headways at peak, achieves at best 24 second dwell times in central section, but 30-40 seconds is more realistic for most stations.
Don’t forget, dwell time includes more than just passengers getting off and on. It has to include time to open the doors, close the doors (including a 2-3 second visual and audible warning!), perform needed safety checks, and eventually pull away. Those operational components the sandwich the core “people getting on and off” bit of station stops add up to a non-trivial number of seconds.
https://www.researchgate.net/figure/Dwell-Time-and-Passenger...
As a minor note, the NYC subway uses a system of fix automatic blocks, not a moving block system, which is what your simulation appears to use. Moving block systems are dramatically more efficient than fixed block systems. But I have no idea how you would get hold of accurate block locations for the NYC subway.
Lines with CBTC use moving blocks. That’s the L, 7, and parts of the E, F, M, and R.
You don’t need a warning for opening the doors. The warning for closing the doors will overlap with people boarding.
That being said 10s is too low as a default but it can easily be changed to a more reasonable 30s or more.
I'm 100% showing this to them today just for fun. They'll get a kick out of it.
Is any of your/their work published?
For example: Selected: Black fill Normal operation: Color fill with 100% opacity Slowing down: 70% opacity Stopped: striped fill, 50% opacity At station: pulsing opacity
A slider to do a bit of time-travelling if possible would be also a nice to have
Sometimes when I Edit Routes and click an "<- Add" button I get the console error "Uncaught Error: coordinates must be an array of two or more positions" and the page blanks out.
-->-->-- --<--<--
I'll add some logic to prevent such options from showing up.
All input data and scripts are in src/data. Run the two Overpass queries, replacing New York with a city of your choice, and you will get JSON data on the infrastructure, as well as services. Replace the two existing JSON files, run the two transformation scripts, and you are done.
Edit: Actually Clifton is in NY, and playing with Google maps there's ZERO public transport from Staten Island to NJ, except by going through Manhattan!
So my idea wouldn't help anyway, unless they extended that subway all the way over the Goethals Bridge.
Forest Ave/Goethals Rd North to Bayway Ave at Mckinley St.
Edit: This could be a fun game - find the two spots with the greatest ratio of driving vs public transport. I think those two are pretty good candidates.
Distance: 2.5 mile.
Drive 9AM: 5-9 minutes, 11PM: 4-6 minutes.
Public transport 9AM: 1:52, 11PM: 3:34.
Ratio (using the average drive time) 9AM: 16, 11PM: 42.8.
https://maps.app.goo.gl/WU1jzjYh8tkTCuh87
According to Google Maps it's not in this case - I'm getting a walking time of 14:37 (!), where the "walk" includes a ferry from Weehawken to Manhattan and the Staten Island Ferry. However Google Maps doesn't seem to know you can walk across the Goethals bridge.