Analysis of 514 Cherry Service for December 2016

The 514 Cherry car has been running since June 2016. Although originally planned as a net new service, budget for the route fell victim to the 2016 round in which headroom for the “new” service was created by reallocating vehicles from 504 King. The purpose was to concentrate service on the central part of King where there is higher demand, but in practice, the original schedule did not work out. In November 2016 the headways on 514 Cherry were widened to compensate for longer-than-planned running times.

The 514 Cherry car has been something of an afterthought for the TTC in several ways. Planning and construction for it began years ago, but implementation was delayed until after the Pan Am Games were out of the way and the Canary District began to populate with residents and students in the new buildings. Another major blow has been the failure to build the Waterfront East LRT which is intended to eventually connect with the trackage on Cherry Street as part of a larger network. In effect, the spur to Distillery Loop is treated by the TTC as little more than a place for a scheduled short turn of the King Street service, much as trackage on Dufferin Street south of King is for the route’s western terminus.

Riders bound for the Distillery District face two challenges. One is that the older streetcars do not have route signs for 514 Cherry, only a small dashboard card wrapped over the “short turn” sign. Tourists might be forgiven for wondering if a 514 Cherry will ever show up. As new streetcars gradually appear on this route, this problem will decline, but it is an indication of the half-hearted way service was introduced that good signage was not part of the scheme.

New low floor cars now operate on 514 Cherry, typically two in off-peak periods and four in the peak. However, the TTC appears to make no attempt to assign these cars to runs that are equally spaced on the route, and so it is common to see both of them near one of the other terminus with a wide gap facing anyone who actually needs to wait for one.

Indeed, it is the same pair of runs that usually have a Flexity on them through much of December, and they do not provide evenly spaced accessible service over the route. The TTC is happy to crow about accessibility, but falls down in the execution.

Worst of all are the actual headways found on 514 Cherry. Although the schedule was revised in November, and cars should generally have time to make their trips, it is very common to see two 514 Cherry cars close together followed by a long gap. This problem originates at the terminals, the points where the TTC’s target for “on time” service is no more than one minute early to five minutes late. This six minute window is routinely broken by service on the route, and the problem only gets worse as cars move across the city.

In effect, the TTC has simply thrown out a bunch of extras for the King car and lets them run more or less at random providing supplementary capacity in the central part of the route.

Headway Reliability

“Headway”, or the spacing between transit vehicles, is key to the provision of reliable service. Would-be riders find unpredictable arrivals among the most annoying “feature” of transit riding, especially when the scheduled service is infrequent to begin with. There is no guarantee that if one just misses a car that another will be along soon.

The following charts show the observed headways northbound at Mill Street (just north of Distillery Loop) and at Springhurst Avenue (just north of Dufferin Loop) for December 2016.

For readers who are unfamiliar with the chart format I have used in these analyses, there are 11 pages divided into sections:

  • The first five pages show the headways for each week in the month, with days of the week shown in separate colours. Each dot represents one vehicle with its horizontal position giving the time, and the vertical position giving the headway in minutes. A trend line is drawn through the points to show how they behave collectively. The degree to which points are scattered above and below the trend line gives an indication of the variability in the values. Note that this is lower for the last week of the month when conditions on the route were more benign.
  • The sixth page shows all of the month’s headway data on one page to give a sense of the overall level of scatter in values. Note that many are above the 20 minute line and some beyond 30.
  • The seventh and eighth pages show weekend data. There is considerably more scatter in the values on Saturday than on Sunday, but in both cases many points lie outside of a six minute window.
  • The last three pages show averages and standard deviations (SD) for the weekday, Saturday and Sunday data. Ideally, the SD values should be low if the TTC is to achieve its reliability target at a value of 2-3 minutes. This would indicate that a substantial portion of the headways lay within 2-3 minutes either side of the average. The actual SD values are generally five minutes or greater.

What these charts show is that the service right at the terminals, the point where TTC measures “reliability”, does not meet their target and misses by a wide margin. This dooms the entire route to uneven service and riders to unpredictable waits for cars. The problem is not confined to a few days, but exists throughout the month.

At the midpoint of the route, Yonge Street, the westbound service is effectively a Dufferin short turn of the King car, and so headways of 514 cars do not really matter. However, eastbound service goes to the Distillery District, and riders might reasonably wait for a 514 to show up rather than taking a 504 King. Sadly, the uneven service leaving Dufferin Loop is as much of a mess when it gets to Yonge Street. Gaps of over 20 minutes are common.

Travel Times

An important issue for any route is the adequacy of the schedule compared with actual conditions of road traffic and passenger loading. The 514 Cherry running times were updated in November 2016 to allow more time for trips across the city. (Click for a larger version.)


The actual running times between Mill Street and Springhurst Avenue are shown in the following charts. Note that this does not include time spent at the terminal loops which is discussed later on.

The travel times do show some variation, but considerably less than the headways. This is not unusual, and is the sort of effect that can be smoothed out at terminals. The week before Christmas is interesting because it shows the effect of pre-holiday traffic building up at midweek and then falling by Friday, December 23.

Thursday, December 15 was a snow day with the effects hitting during the PM peak.

If running times were very tight, one would expect to see a lot of short turning and/or very short layover times at terminals. However, the actual time taken at the terminals indicates that 514 Cherry cars generally had comfortable schedules that allowed a layover.

These charts show the round trip time from intersections just outside of the terminal loops and include any layovers the cars might take. Some of these are quite short (showing a car that basically ran through the loop with a minimal layover) while others, notably on weekends, were quite long. The irregularity of headways leaving the terminals cannot be put down simply to cars always being late, but to operators taking a leisurely break. Note that there is a staff washroom only at Dufferin Loop, and so there is an incentive to take longer layovers at that end of the route.

One Day’s Operation In Detail: Thursday, December 1, 2016

This section reviews one day’s operation at the detailed level. First is a chart showing the movement of all vehicles.

This set of eight charts shows movement of vehicles on the route over the entire day. The east end of the line (Distillery Loop) is at the bottom, and the west end (Dufferin Loop) is at the top. Each page covers 3 hours’ operation from left to right. Each car is represented by one line sloping to the right upwards (westbound) or down (eastbound). A car that is stationary shows up as a horizontal line, typically at terminals and at intersections. A few notes:

  • GPS errors can cause a car to reverse direction in the plot or even disappear (typically because the reported locations are consistently well off of the route for several minutes ). Some of these are “ironed out” either by dropping rogue data or by interpolation over gaps, but there is a trade-off between “cleansing” the data and producing erroneous plots. The result is that some oddities remain in the charts.
  • Congestion shows up as a change in the slope of a line, typically near a busy location.
    • This first appears westbound toward University Avenue just at 9 am where streetcars, notably the one charted in turquoise, spend time pulling up to the stop bit by bit as traffic allows. Examination of tracking data for 504 King shows that a parade of 504 cars arrived at this time, and the 514 was delayed as each car in the pack moved up to the intersection, loaded and then departed. This was a problem of streetcar congestion, not of traffic.
    • Congestion also shows up in both directions between Bathurst and Spadina beginning about 3:30 pm. One problem facing 514 Cherry cars was the traffic backlog eastbound at Spadina caused by 501 Queen cars making the east-to-north turn without any transit priority signal assistance.
  • The length of terminal layovers is visible in the width of horizontal lines for vehicles at Distillery and Dufferin Loops.
  • A group of three cars forms up westbound at Bathurst just before 4:00 pm. These travel together until about 6:30 pm when some of the problem is sorted out with a few short turns at Spadina, one in each direction.
  • Service remains irregular throughout the evening.

The service can also be viewed by looking at headways at various points along the route.

Each page of the headway charts shows data for one location working across the route. The values plotted here for December 1 are the same data that appear in the monthly charts above for that date.

  • Westbound headways at Mill are reasonably well-behaved through the AM peak with one long gap at around 10:00. However, beginning at about 1:00 pm, the headways start to be erratic with wide swings showing pairs of cars travelling together were common (a high value immediately followed by a low one).
  • Stepping from page to page, one can see how the pattern leaving the terminal travels across the route. The morning service that was well-spaced leaving the Distillery becomes less so as it travels across the city, but the condition of afternoon and evening service is set right from the terminal.
  • The same patterns are evident eastbound from Dufferin Loop with afternoon and evening headways at Springhurst showing wide swings and pairs or triplets of cars. One wonders why anyone would wait for a service that appears on a headway over half an hour long.

Travel times can be broken down by route segment to reveal where there are problems with congestion or other chronic delays.

In these charts, each dot represents one vehicle travelling between the two end points of the segment. In some segments, the times vary little all day long, while in others there is a clear rise and fall due to congestion and/or passenger service times.

  • Westbound from Yonge to University at the end of the AM peak. Some of this is a result of the 504 King parade mentioned above.
  • Westbound from John to Spadina at the end of the PM peak.
  • Westbound from Spadina to Bathurst through the PM peak and early evening.
  • Eastbound from Bathurst to Spadina in the evening.

Some of these effects are specific to individual days, while others have day-of-week and seasonal effects, notably congestion in the Entertainment District.


Although 514 Cherry is a “new route”, it is little more than a “scheduled” short turn of 504 King, and that schedule bears only passing relationship to the service actually provided. For those attempting to travel to or from the Distillery District, the service is unreliable, and the primary cause appears to be a lack of line management and headway discipline, not congestion along the route.

10 thoughts on “Analysis of 514 Cherry Service for December 2016

  1. Another challenge 514 users have is that in the mornings (before 7am) and at night (after 10pm) the streetcars do not go down Cherry, but instead continue on King to loop around (Broadview-Dundas-Parliament-King), and a shuttle bus runs from Cherry to Parliament (I see that these times are not represented on your charts). This arrangement is to mitigate the alleged excessive noise of the streetcars in the turns at Distillery loop and and King/Cherry. My spouse relies on this line to commute to work, but the shuttle bus is *extremely* unreliable, and the arrangement means that a further transfer is required. Any analysis of actual service levels and reliability would ideally include these periods if the data are available.

    In community meetings, the TTC has suggested that this situation may be resolved when enough Flexitys are available to cover the full line (they supposedly have better track lubrication which reduces the squeal). In the meantime, from discussions my spouse has had with the shuttle drivers, it appears they are getting time-and-a-half, in addition to the regular 514 drivers, and riders are getting poorer service. (I think it would likely be far cheaper to just offer a portion of what the TTC is spending on the shuttle service to the very few people who are loudly complaining…)

    Steve: There is no tracking data for the shuttles, and so nothing to include in the analysis. Sorry.


  2. The area around the Ontario street stop has changed rapidly in the past few years, and I doubt the TTC is ready for this. The new Globe and Mail Centre has only a few tenants moved in and yet this morning I could see there’s already a dozen people exiting every eastbound 504/514 there to enter the building at the peak of rush hour. I expect that to go up significantly, despite their fancy free shuttle to Union. George Brown students also arrive at that stop, and nearby stops in waves, many heading to/from the new residence building by Cherry. The service needs to be really reliable to handle these crush events, and it isn’t.

    Steve: At some point, I will publish an analysis of the combined 504/514 services on King Street. This article was intended mainly to show the utter lack of reliability in terminal departure times, the metric near and dear to the TTC’s heart.


  3. You say: “…the service is unreliable, and the primary cause appears to be a lack of line management and headway discipline, not congestion along the route.”

    Really rather sad but based on your exhaustive (and exhausting!) analyses you could probably say this for almost all (or all?) routes. Do you know if your blog is actually read by people at the TTC who could, if they tried, do something to change this lackadaisical ‘management style’??

    Steve: Yes, it is.


  4. I took the 514 on New Years Eve from the Dufferin Gates. The service was prompt that one time.

    I took it on Labour Day AFTER the parade and despite the published information, the streetcar did not return to the gates until much after the parade finished.

    Personally I think the TTC would have been better off to run this as a scheduled turnback service rather than a separate route. I always believed the creation of the 514 was a PR stunt designed to make it look like they were providing more service when in reality they wanted to try and eliminate issues on the 504.

    On paper it helps eliminate issues at Dundas West and Broadview and provide more service through Downtown. In reality, it is just another 504 Streetcar.

    Just curious Steve, any interest in doing a comparison on congestion at Dundas West and Broadview pre and post 514?

    Steve: I can have a look at it, but from my own observation, the congestion is still with us. It’s a matter of degree.


  5. Are you aware if signal priority has been implemented at the King and Sumach intersection for the 514?

    Steve: Sort of, but my sense from riding through this location regularly is that it could be improved. When a car is northbound on Sumach, or eastbound on King with the switch set to turn south, there is a transit white bar phase. There have been times when I have seen this with no streetcars in sight, but not recently, and so it may actually be triggered by transit vehicles now. When the line opened, it was just part of the regular cycle. Of course, this location also has a permanent slow order for noise control, but that’s not part of the signalling issues.


  6. Is 509 all Flexity now? How long before 509 and 514 conversion to Flexity complete? We have all got to thank AODA for it without which we would still be riding inaccessible frozen old streetcars.

    Steve: No, it’s not. There are typically two cars on the 514 line at midday, up to four at peak. Complete conversion depends on Bombardier’s delivery performance, but I would be surprised to see it before late Spring 2017 at best.


  7. L Ball says: “There are new Nextbus digital screens in the shelters as at EB Yonge. That’s gotta count something at least.”

    Indeed! This display was out of action for about a year and after lots of back and forth with TTC it seems that Hydro had cut the power. They finally restored it but then put it on same schedule as the street lighting (i.e. it only worked at night!). After TTC were unable to get them to do it properly the TTC tried a solar powered display as a test and it seems to work. I understand that they are looking at buying more of these for other shelters without power or with Hydro problems.


  8. Steve wrote: “The TTC is happy to crow about accessibility, but falls down in the execution.”

    This is a trend in many areas. One of the most disturbing of these areas can be seen in some of the signs posted by elevators that are out of service for maintenance. These signs indicate the alternate route to use since the elevators is out of service. Many of these alternate routes are profoundly convoluted, bizarre and involve serious detours out of the way. My joke is, “And the next instruction on the sign is ‘… then take a dogsled to Nunavut.'”


  9. The area around the Ontario street stop has changed rapidly in the past few years, and I doubt the TTC is ready for this. The new Globe and Mail Centre has only a few tenants moved in and yet this morning I could see there’s already a dozen people exiting every eastbound 504/514 there to enter the building at the peak of rush hour. I expect that to go up significantly, despite their fancy free shuttle to Union. George Brown students also arrive at that stop, and nearby stops in waves, many heading to/from the new residence building by Cherry. The service needs to be really reliable to handle these crush events, and it isn’t.

    This is why many such as myself were perplexed by Planning’s decision making on the RL where they seemed to be chasing lofty goals like the hypothetical development of Moss Park and justification of fulfilling the greater good downtown (by providing theoretically “better” coverage via Queen). Prioritization of the needs of actual riders? You expect too much.

    I happened to be in the area very early in the morning this past weekend and I too noticed bodies leaving streetcars and entering the office buildings. I can only imagine the nightmare of pedestrian congestion on the sidewalks on the few north to south streets when the office space is fully built out if the main RT line into the area is well north of the office buildings.

    The best part is the 351 King shuttle someone brought up previously. It sits in the outside lane blocking it entirely while taking layovers.

    Steve: I believe that the choice of Queen is political to avoid having the RL too close to SmartTrack on the map. There are utility and geotech issues with King, but it’s so much better for a western extension.


Comments are closed.