The information in this post just appeared on Facebook, and so I no longer feel bound to treat it as “pre release”.
The NextBus TTC displays now include the St. Clair, Bathurst, Carlton, Dundas and King routes in addition to Spadina and Harbourfront which have been publicly visible for some time.
Real time predictions for next vehicle arrivals are available as well as maps showing vehicle locations and movement.
Now you can monitor the quality of service from the comfort of your own computer. Warning! Some images may be unsettling, such as the view I had earlier today when all but two St. Clair cars were between Oakwood and Bathurst.
The NextBus site works with PDAs and will detect the type of device you are using. This access has its own URL, although you can navigate down simply by going to the home page at nextbus.com and drilling down to the TTC from the selection of locations and systems. Maps on PDAs are available, but they don’t refresh automatically, and they are much simpler as befits the small screen.
Gotta love the Kwality Assurance. Route 506, “Calton”?
Steve: They did say it was a “beta” version.
Thanks for sharing this Steve! Just curious, does Nextbus have live maps for subways and higher order rapid transit too?
I’ve seen “Carlton” spelled “Carleton” lots of times. This is understandable as it is also the spelling of Guy Carleton, Baron of Dorchester; the namesake of Carleton County (now vanished into Ottawa) and the university there. But “Calton”?
Still, an interesting link. Would the 501/502/503 be too embarrassing to add yet?
I assume this means we’re now just a flick of a switch away from getting this information displayed on those digital displays that have been installed at a smattering of shelters.
Or, in other words, four months away.
I noticed they use > and < symbols. WAAAAAY better than the GPS map at Spadina station.
Have you seen that GPS map at Spadina? you get like 3-5 streetcar symbols together, you don't know if it's Southbound or Northbound.
Is it me or does the live map option on a pc takes way too long (internet time)?
Anyways between my last sentence/question I tried the nextbus on my blackberry pearl and it looked nice.
I like that I get an option of next time and a live map. Some people will want to know the time (text) and some might want the map itself.
Will nextbus be integrated to the TTC's trip planner?
How come they didn't use myttc's data/work/whatever?
I see 504 has that dundas west/lansdowne/queen diversion.
There is a southbound ttc vehicle north of Queen. Interesting.
Does this really show all the cars on the route?
I’m looking at the 512 right now (9pm, Monday) and there are four cars on the line. According to the schedule on the TTC site they should be on a 5:30 headway but for the approximately 10km round trip, that would mean they’d be averaging close to 30km/h, including stops for passengers, lights and the three loops. Are the schedule makers that off-the-wall, is the line a car or two short right now or is one or two not showing up in the system?
Obviously they aren’t close to a 5:30 headway (doesn’t help that right now one is following another only about 3 minutes behind, leaving a 15 minute gap behind him).
Steve: Later at night (11 pm as I write this), Nextbus shows only one car on the line.
The map will appear rather disjointed when they show a 506 CALTON down in Coxwell-Queen Loop! Or perhaps those short-turn cars fall off the edge of the planet, as I suspect the Nextbus technology is about as modern as the thinking of the Flat Earth Society.
St Clair has only one streetcar on it according to Nextbus, 35+ minute gap (10:35 Monday night). How complete is their dataset?
Is it normal for there to be only 1 streetcar on St. Clair at 10:50pm tonight with a 30 to 40 minute headway?!
Steve: No. Rather strange. Well they did say it’s a beta system.
I noticed on the weekend that the eastbound King car stop at Spadina had an electronic next bus stop installed. It just kept saying “Next vehicle information, the time is…”, but it looks like maybe a rollout is imminent?
Steve: Yes, I believe it is.
Steve: “Warning! Some images may be unsettling”
Like the view I’m getting now that shows only one streetcar on the entire St. Clair route with a 40 min headway. I know it’s late, but only one car seems a bit thin for a route they just spent so much effort on. Every other route shown has more service (Bathurst, which seems to be of similar length has 4 providing roughly a 10 min headway).
I’m sure that if given the proper nudging you someone will manage to get a screen set up at surface stops at stations which don’t have a fare paid zone using wifi access from coffee shops which seem to be located at these locations. Yes, I know the TTC is planning to roll out proper signage to do the same thing, but given the chance I think someone will try to jump the gun given the speed at which such projects get completed. It’d be a nice example of grassroots local cooperation.
Steve: I can think of all sorts of uses for this information including TTC’s own route supervisors actually being able to see where vehicles are.
Google Map takes forever to initialize and according to the website no SMS support for Canada yet. Also similar to Joey I see a 39 minute wait for 512 Westbound St. Clair/Vaughan at 12:20am when schedule shows 8 minute intervals. Something is amiss. Any chance the concept of “sunlight as the best disinfectant” will apply here and Nextbus scrutiny from the general public will drive service improvements?
Steve: Leaving aside any problems with the data feeds, I think that the availability of this info will allow widespread view of typical service, and will leave TTC management with nowhere to hide on quality issues.
I am watching 3 westbound cars at oakwood, now reaching dufferin.
The nearest next westbound car is at Spadina.
I give up. Maybe the TTC should be banned from running streetcars since they can’t seem to do it without screwing it up.
Steve: I’m looking at the route now (12:30 pm), and the service looks to be in good shape. One of the reasons the historical reviews I do are important is that we can see how bunches and gaps develop, and whether this was due to a delay or lack of attention to spacing service properly. All the same, the ability to look at any route in real time will make it much easier to see how much attention, or not, is being paid to service regularity.
John F Bromley says:
March 8, 2010 at 9:53 pm
It shows the 519 and 509 cars sitting in the car house as soon as the run number information is entered into it. It then shows the cars as they come along King, Queen and Bathurst to get into service. It also show them on detour.
Steve: I presume you mean 511 or 510, not 519, or does the TTC have a new route they have not yet advertised?
After checking the 509 and 510 information over the past week and now inspecting the newly added routes, I have not been able to view the maps — in various browsers.
I’d like to see some objective studies of the accuracy of these arrival time estimates. Recent messages from transit riders in York Region have called into question the usefulness of the Viva timers. These have been in place for several years.
My personal experience with the subway monitors is that the predictions are fairly accurate. However it is during delays that good information is needed … and I don’t have a picture of how well the monitors are serving customers then.
Steve: As I write this, I have a NextBus map up in IE 8, Safari, Firefox and Chrome, and I use it regularly on my Blackberry Curve. A friend has forwarded screenshots of it from his iPhone. You need to have Java enabled.
The subway “next train” times come from track circuit occupancy. For example, if there is a train sitting at Pape, the sign at Broadview says “3 minutes”. Once it leaves, this changes to “2 minutes” and stays there until the train leaves Chester which triggers the “1 minute” notice. Inaccuracy arises wen there is a delay or layover caused by crew changes (westbound at Coxwell, if that’s the next train east of Broadview) or by holds for spacing (westbound at Chester, commonly).
On the surface system, I think NextBus is still tweaking its prediction process so that a car at location “x” is reliablity a certain time away from downstream locations “a, b, c, etc”. A related problem is with short turns that are “off route”, and therefore not part of the prediction scheme, until they come back into service. A Carlton car going to Bathurst Station is a good example. NextBus has no way of knowing if or when it will become an eastbound car on College until it actually shows up there.
Hmmm… “wireless” (PDA) versions offer three different resolutions (good) – except the medium and large are just stretched versions of the smallest, so it looks all pixelated (bad).
Steve: And there does not appear to be a “zoom” function allowing one to look at a smaller portion of a long route in detail. That’s what beta testing is for — find the bugs and the enhancements people want.
This really is an elegant program. Yes, it’s still in beta, and clearly not every streetcar has been fitted with a device that contributes to the data, but what I’m seeing here tells me that I’m going to find this service extremely useful in the near future. I’m picturing pulling this stuff up on an iPhone, and finding out when the next cars are due, just as I come to a stop.
You know what would be a good idea? They’ve already made it possible for you to pay for goods using your iPhone (you pass an on-screen barcode through a laser reader). Imagine a service where you pay for a day pass on the iPhone, and then just wave the iPhone through a reader as you board a streetcar or bus.
The future really isn’t that far away, if we work towards it.
Chicago’s transit authority (who seem wondefully clued-in about the advnatages of the web for transit) recently produced an on-line “DIY Bus Tracker Display“. The idea is that any internet-connected computer can show the upcoming buses from any location – one can imagine it in the lobby of a big office or condo building, or a shop window, or in a mall. The easy configuration tool can get you something usable in about 30-40 seconds.
Somehow, Toronto feels rather behind the times…
Steve: I think some parts of the TTC only dimly understand the power and effect of making access to their data transparent and friendly. For far too long, they regarded it as “proprietary” (see also the fight about subway station buttons) and failed to recognize the marketing and customer service value of giving the info away.
Steve: “I presume you mean 511 or 510, not 519, or does the TTC have a new route they have not yet advertised?”
Yes. I suffer from hunt and peck typing and lack of reading glasses.
I suggest the pointers for cars in opposing directions on a route be in different colours. This would make it easier to isolate cars traveling in the direction one is interested in monitoring.
Could cars being short-turned have a special indicator – fluorescent label, flashing etc.? Why go to the corner to catch a car that won’t travel to the end of the route?
Perhaps each route/direction could have its own colour label? 504 w/b, green on white; 504 e/b white on green; 508 w/b, blue on white; 508 e/b, white on blue?
Did they take down the additional routes? Yesterday it was showing Bathurst and St. Clair routes and now it seems to only show Spadina (510) and Harbourfront (509).
Steve: Yup, they’re gone again.
Are the time announcements for VIVA, based on a schedule or on the actual positions of buses?
Now same question for TTC buses/streetcars? MANY MANY MANY routes are known to never EVER stick to schedule, so if they install NEXT VEHICLE based on schedule instead of REAL GPS then they would be useless.
I seen the NEXT TRAIN IN screen at Warden station (for westbound) say 7 minutes then go down to 4 minutes. Do they skip minutes? they don’t seem to be on real time where the trains are. many times, the train hasn’t left the station and the time for the next train is on the screen.
These NEXT IN signs need to be using REAL GPS and not based on schedule.
Real GPS = stick a gps tracking unit on the buses that sends a magical signal to outer space then it goes back to wherever operations HQ is and they send it to the machines on stops. NOT BASED ON ROTATIONS OF THE WHEELS.
Steve: The NextBus prediction is based on actual vehicle locations.
I think it should also be acknowledged that the interface on this website is just plain awful.
This morning, they had 504 (showing the diversion, but no buses), 505, 506, 511 and 512 in addition to 509 and 510. Now, however, they’re just back to 509 and 510.
It would be great if the destinations of the streetcars could be indicated. Maybe the car labels could replicate what’s shown on the rollsigns of the cars.
Steve: I can see this sort of thing once the streetcars have digital signs and assuming that the code for the current display is included as part of the status transmission from the car. Whether the TTC had the smarts to include this in the work on retrofitting buses with GPS I do not know, but suspect not.
I have a few questions for the TTC about this data as well:
1) Is there other vehicle data being collected: doors open/closed, gas, driver id, sign info, vehicle orientation (north, south etc.), # fares collected
Steve: There is some additional data, but little from the list here. The driver id is present, although I don’t ask for it in the data extracts I get as I don’t want to be in a position of trying to track chronic “bad actors”. If the TTC actually analyzed their own data, they could do that sort of thing. My interest is in the overall quality of service.
2) Is there information about whether the vehicle is on it’s way to a route or on route, and if it is in service (ie. it may just be driving to the route but not picking anyone up enroute).
Steve: This depends on how the operator has “signed on” the vehicle to the network. As you know from the data you have modelled, you can usually see cars sitting in and then leaving a carhouse, as well as much of the return trip. However, running express or out of service is not a status the system knows about. As I have written before, the TTC should be capturing the vehicle destination sign code as part of their data stream and making the meaning of the codes public. However, right now, assignment of the codes is something between the Planning Department and individual garages. If this data takes on a separate value for purpose of vehicle tracking, then there will have to be better co-ordination of changes in the codes on a system-wide basis.
3) In the data I got from Steve the data was every 20 seconds, is it possible to do 5 or 10 seconds (this would help with identification of delays at stoplights etc.)
Steve: This is a constraint of the CIS system. It polls every vehicle once per 20 seconds. It is actually possible to see non-trivial stoplight delays in the current data streams, and I plan to analyze this in more detail to determine whether there have been changes in behaviour at specific intersections.
4) And the obvious one … when is historical data going to be public, and when is NextBus going to open up a public API?
Steve: Several people are hammering away at the TTC about making the raw data public both for historical and real-time access. NextBus does not want to give away their prediction info as it is their intellectual property. The raw data, however, should be public as-of-right.
Quick question: On the TTC’s website, it tells you the next 3 vehicles at scheduled stops. Are these times updated in real time, even if just on certain routes? Yesterday I took the 53 Steeles East bus between a stop near Don Mills and Finch Station, and both ways they were virtually exact to the times on the website. In fact, coming back the schedule posted at Finch Station listed different times than the website, and the bus was clockwork on the website’s time.
Steve: Both the TTC website and the information on pole cards is taken from the schedules. However, the website has current info while the pole cards may be antiques.