Updated at 3:20 pm, December 27: The scale on the headway charts has been changed to 30 minutes with 3-minute gridlines, and on the link time charts to 18 minutes with 3-minute gridlines. The intent is to spread out the data points to give a better view of the fine details.
[My apologies for the appearance of this series many months after the fact. It took quite a while to get the GPS-f0rmat data from the TTC for reasons unknown.
I look forward to the plans for an Open Data access to data for all routes as a regular online service so that this type of analysis will not require special requests for data extracts. Whether the “new” TTC goes ahead with providing this data remains to be seen, although “transparency” is supposed to be a watchword in the new administration.]
June 30, 2010, brought streetcar service to the full St. Clair route out to Gunn’s Loop for the first time since 2007. In an earlier article, I reviewed the line’s operation in April 2007. This was a “before” snapshot intended as a comparison to the “after” construction line behaviour for which we have waited so long. Now we can look at the “new” St. Clair to see the benefits, such as they might be.
On two weekends, service did not operate over the entire line due to street festivals. No sooner was streetcar service restored to Keele, but it vanished again on July 3-4, and again on July 17-18.
This article reviews the basic information available and some of the analysis I have done using a single weekday as an example.Overview of Service
July 5, 2010, was the first day of “regular” weekday service after the Canada Day weekend. The chart linked below tracks the movement of all vehicles on the route.
This format of chart dates back to the late 1800s when it was invented in France for building railway schedules, and it is still used today for that purpose.
Time flows from left to right beginning at 4:00 am and running over 8 pages to 4:00 am the following day. (Night service on St. Clair is provided by a bus that runs with a different route number, and even if the buses were sending GPS-based data, it is not included in an extract for route 512. A similar problem afflicts the NextBus system which does not integrate daytime and overnight routes as one entity.)
Distance flows on the vertical axis with St. Clair Station (the eastern terminal) at the bottom and Gunn’s Loop (the western terminal) at the top. Major intersections are noted by lines across the chart.
All positioning data are GPS based reported every 20 seconds from the vehicles. Wiggles in the lines show the variation in speed, and horizontal stretches indicate that a vehicle is stopped. This is notable at the termini and at St. Clair West Station, but also for shorter periods at intersections where it is possible to distinguish the nearside and farside stops. (Occasionally, a vehicle reports a rogue location from its GPS. These are usually trapped out in the data mapping process because they are not reasonable locations for a vehicle on the route. Sometimes, an erroneous location will appear that is near enough to the route to appear as “real” and this can cause a larger wobble in the chart line for the affected vehicle.)
At St. Clair West Station, westbound cars appear to change direction, but this is caused by their moving around the loop to the exit track which brings them back slightly east. At St. Clair Station, the majority of the layover time is taken on the offloading platform and cars move to the loading platform for a comparatively brief period before circling through the loop and back out onto the street. The very bottom of the chart is the east end of the loop.
Short turns are easily seen as cars that do not reach the termini, but sit at an intermediate location waiting to re-enter service. For example, just after 6:00 am, a car (dark orange line) short turns at Earlscourt Loop (Lansdowne). This was not a scheduled move, but I suspect that the car entered service late from Roncesvalles Carhouse and is getting back into position. (Looking at other days, this is a fairly common event.)
There are no scheduled short turns on the 512 route now and all service is supposed to operate from Yonge to Keele. Before the line was rebuilt, there was a scheduled short turn service to Lansdowne in the AM peak period.
At about 10:20 am, service was blocked eastbound at about Old Weston Road. This cleared at about 11:05. Construction was still underway at this location, and this may have contributed to problems at the intersection. The first car (green line) to be held went out of service eastbound suggesting that it was involved in a collision. The other three cars were short turned at various places to get back on time. All other cars short turned at Lansdowne to maintain service eastbound.
Bunching is evident in various places.
- At times, we can see a slow car with its follower catching up. For example, the red and mauve lines westbound from Yonge just after 13:00 show two cars converging, but they re-establish their headway eastbound from Gunn’s Loop.
- Conversely, some cars run for extended periods as if they were coupled together. Just before 15:40, a pair of cars (one of which has just entered service) heads east from Gunn’s Loop. They stay close together until about 18:10 eastbound at St. Clair West Station.
- Around 17:40, one car (dark blue) makes a round trip during which it is gradually joined by not one but two following cars (pale and dark turquoise respectively). The headways during this period are quite erratic thanks to this bunching which has no apparent cause. Indeed, there is more reliable service late at night (after 22:00) than during the early evening, although some bunching does occur caused by extended layovers, not by that old TTC bugbear “traffic congestion”.
Of particular interest given the stated goals of the St. Clair right-of-way is the ability of service to operate reliably over the route without interference from traffic. Variations in running time should arise mainly from boarding delays at heavy stops rather than from traffic interference.
The next pair of charts shows the times required to traverse each section of the route.
The location “St. Clair Station” is east of the off-loading platform. Therefore, the link westbound from the “station” to Yonge Street represents the departure trip, while the eastbound link includes the layover at the platform.
The westbound links are generally uneventful, and stay fairly well clustered around the trend line showing that running times are quite consistent from car to car. Over the day, there is little variation in the trend line indicating that peak period effects are quite small. There is somewhat more scatter in the data for the link through St. Clair West Station (Tweedsmuir to Bathurst) showing the variation in dwell times at that location. One outlier at about 21:40 is a car that took a long layover in the station and then re-entered service.
The eastbound links show the effect of the delay at Old Weston Road. The first car was held in the Keele to Caledonia segment, while the following three remained west of Keele, and their delay shows up in the Gunn’s to Keele segment. Within the Caledonia to Landsowne segment, we see the effect of crew changes at Lansdowne, a few of which hold cars for extended periods. The Yonge to St. Clair Station segment shows the size of and variation in layovers taken at this terminal.
Overall, there is little visible effect from the peak period on running times over the entire route. In a later post, I will review this data looking at the month as a whole, and at the route as a whole.
Although the link times show no pervasive effects from traffic congestion or loading delays at any time of day, the picture for headways is far from ideal. The scheduled headways on St. Clair are:
- AM Peak: 2’55”
- Daytime: 6’00”
- PM Peak 3’30”
- Early Evening: 6’00”
- Late Evening: 9’00”
A common situation in all of the headway charts is the spikiness of the line tracing the data. Although the trend line follows the overall pattern of the scheduled headways, the individual values wander quite a bit from the target. By TTC standards, a vehicle can be up to 3 minutes off schedule and be considered “on time”. This produces the absurd effect on a route like St. Clair where pairs of cars can operate together (one 3 minutes ahead, one 3 minutes behind schedule), providing much worse than the advertised service, and still be considered “on time”. Statistics purporting to show service quality using this type of metric are meaningless.
The effect of the delay at the west end of the route is obvious at Yonge Street in the late morning and midday when headways widen due to short-turns further west. Of particular note is the large gap at about 12:30 over an hour after the delay at Old Weston Road cleared. Looking at the chart of all service, one can see that short turning to sort out the headways didn’t actually begin until about 12:10.
Bunching is common through the PM peak and early evening causing headways well above the advertised level.
The eastbound headway charts show similar behaviour.
What is quite striking here and in charts for later days in the month is the degree to which headway regularity is not a hallmark of the operation even though the major source of delay has been eliminated with the right-of-way. I will return to this in the next article.