In the previous post of this series, I looked at headways along the Dufferin route for December 2006. Now, I will turn to the Link Times, the length of time taken to get from one point on the route to another.
If these times are well behaved, this indicates that the requirement for a bus to cover this particular section is predictable even though it may vary over the course of a day or by day of the week. Random interruptions occur rarely and the schedule can reliably make assumptions about travel times.
If Link Times are spread out over a wide range of values, particularly for trips at similar times of the day, then something in this area is making travel unpredictable, likely some form of congestion.
When we are considering the reliability of a service and how it might be improved, areas and times with widely varying link times are a good place to start looking. Conversely, if link times for a route are generally well-behaved, then variations in headways have some other cause than random interference from traffic.
Northbound from King to Wilson
As I have mentioned previously, due to CIS data problems the departure time from terminals is not always accurate, and therefore this analysis starts at the first major street north of Dufferin Loop and ends at Wilson and Dufferin, just before Wilson Station. I will return to operations at the terminals later in this post.
29 Dufferin Northbound King to Dundas
29 Dufferin Northbound Dundas to Dupont
29 Dufferin Northbound Dupont to Eglinton
29 Dufferin Northbound Eglinton to Lawrence
29 Dufferin Northbound Lawrence to Wilson
The section from King to Dundas shows quite reliable link times over the entire month with a very small rise in times during the afternoon. One notable exception, as we have seen on many other occasions, is the Friday, December 22 effect which raises the trend line from around 5 to 8 minutes during hte early afternoon peak. We will see this effect with much more serious impacts elsewhere on the route.
From Dundas to Dupont, the link times are more scattered, and we can see a regular rise for the afternoon peak. Note that this is not a sharp peak, but a gradual rise starting about noon, peaking between 1600 and 1700, and gradually falling off through the evening. Similarly, the widened scatter of link times lasts over a long period, not just for a few hours in the traditional peak.
Pre-Christmas effects are noticeable in the week of December 18 with a particularly high peak, around 1530 on the 22nd. There is even some peak effect during Christmas week when traffic is generally quiet.
On weekends, the peak comes earlier in the day corresponding to shopping congestion through this section. A noticeable exception is Christmas itself (on the Sundays chart) when the link time is unchanged all day.
Dupont to Eglinton shows less variation in link times through the week. December 22 produced less change in the trend line compared to other days, but the values are more scattered showing some effects of congestion. On Saturdays, some congestion effects are visible in the scatter of the data, but interestingly the Boxing Day data is quite flat. This is not the case further north.
Eglinton to Lawrence shows the familiar pattern with a slight effect for the PM peak and a noticeable impact on December 22.
Lawrence to Wilson includes the stretch of Dufferin passing Yorkdale Shopping Centre. Here we can see situations quite unlike the rest of the route.
- PM peak congestion is sensitive to the day of the week with longer link times late in the week.
- The December 22 peak is larger than on other links for this route, but the real stunner comes on Boxing Day when the trend line hits 25 minutes around 1300 and individual trips take close to 40 minutes to cover this section.
- Congestion at Yorkdale continues through Christmas week, and the PM peak comes earlier, between 1500 and 1600.
Southbound from Wilson to King
29 Dufferin Southbound Wilson to Lawrence
29 Dufferin Southbound Lawrence to Eglinton
29 Dufferin Southbound Eglinton to Dupont
29 Dufferin Southbound Dupont to Dundas
29 Dufferin Southbound Dundas to King
From Wilson to Lawrence, we see similar effects due to Yorkdale but at a smaller scale. From Lawrence to Eglinton, there is little variation in link times through the month.
From Eglinton to Dupont, times are fairly consistent, but more scattered indicating some congestion effects. Also, this is a rare case, for Dufferin, where we see some AM peak increase in times. It is impossible to tell whether this is due to traffic congestion or loading delays.
From Dupont to Dundas, peak effects are weak, although we see the effect of pre-Christmas shopping congestion. From Dundas to King, link times are flat as they were northbound.
Although CIS does not reliably tell us departure times from the terminals, we can get a sense of operations there by looking at the round trip times from a point nearby. At the south end, the link is from King to Dufferin Loop, while at the north end, it is from Wilson and Dufferin to Wilson Station. Large variations in these times are unlikely to be due to congestion because the links are short and neither is noted for serious traffic delays.
At Dufferin Loop, there is fair amount of scatter in the round trip times from King Street with most values on weekdays ranging from 4 to 13 minutes. The trend lines are fairly flat through the day although there are a few exceptions. Longer times in this chart indicate longer layovers at Dufferin Loop. On weekends, the times spread out indicating that schedules are somewhat more generous, and on Christmas Day the times are roughly double the normal value.
At Wilson Terminal, the situation is similar to that at Dufferin Loop with weekday times ranging over a band about eight minutes wide. There is some variation by time of day, and during Christmas Week, times drop through the afternoon during the period of congestion at Yorkdale we saw earlier indicating that layover times are reduced or eliminated for these trips. On weekends, the scatter in times is much greater and the same bulge with generous layovers appears on Christmas Day.
Probably the most striking part about the Dufferin route is that many congestion effects do not occur at traditional locations or times of the day. Peak period parking controls do little for shopping congestion which is neither directional, in the conventional sense of inbound or outbound peaks, nor timed at traditional rush hours.
The TTC needs to examine routes like Dufferin and major traffic generators like Yorkdale for transit priority that is not designed around standard peak periods.
As on other routes, we see very severe effects from the “Christmas Eve” early rush hour that arrived when roads still operate with off-peak capacity due to parking. This effect should be planned for in the special schedules operated on those days, although the pattern will vary depending on the day of the week that Christmas falls. In 2006, Christmas was a Monday and the early rush hour appears to have co-incided with last-minute shopping. In 2007, Christmas fell on Tuesday, and the shopping was spread out over a longer period from the last working day for many people.
Santa comes only once a year, and there are few opportunities to plan for early rush hours. The next Monday Christmas is in 2017.