Early in 2007 when I started looking at the TTC’s vehicle monitoring (CIS) data, I thought to be finished with it long ago, to have blazed through many routes and written wonderful commentaries on all of them. Things didn’t quite work out as I had planned, and I got bogged down with competing issues and other calls on my time. Also, the programs that digest, massage, and otherwise render presentable the TTC’s data needed some housecleaning both to make them more robust and to reduce a lot of the manual work that went into the early analyses on 504 King.
Things are much simpler now, although the challenges of interpreting the data remain with each route offering its own peculiarities. Now I turn to the Dufferin Bus, a frequent route for which the TTC receives many complaints about service. How will it compare to routes we have seen already?
The route is 13.56km from Dufferin Loop to Wilson Station, although half of the scheduled peak service runs only to Tycos Drive about 3/4 of the way to the north end of the line. This is in the same range as the Carlton and King cars, although they spend much more time in “downtown” conditions. It is shorter than the 16.65km Queen-Humber route, and of course much shorter than the 24.43km Queen-Long Branch route.
The scheduled service is generally more frequent than on the streetcar lines, although with smaller vehicles so that headways are better for any level of demand scaled to capacity.
As I have done on previous routes, I will look first at the data for Christmas Day 2006 as this shows the route in its simplest state without any effects from traffic congestion, weather or heavy passenger loads.
29 Dufferin Service Chart December 25, 2006
29 Dufferin Northbound Headways December 25, 2006
29 Dufferin Southbound Headways December 25, 2006
29 Dufferin Northbound Link Times December 25, 2006
29 Dufferin Southbound Link Times December 25, 2006
The service chart is in the familiar pattern, although there are peculiarities that show up on all of the charts for Dufferin. Note the sawtooth patterns at about 500 (Lawrence Ave.), 340 (Rogers Rd.), 200 (Bloor St.) and 100 (Queen St.). These are all CIS signpost locations, and quite commonly the position calculated from a vehicle’s odometer must be reset when it reaches these points. The buses have not actually moved backwards, only the position used by CIS.
At the top of the chart, you will occasionally see small “tails” extending above the 640 line (Wilson Stn.). These are buses that reported a location to or from Wilson Garage.
Between Dufferin Loop (50) and King St. (80), some vehicles appear to move very slowly. Looking at this closely, it is often the same vehicle on repeated trips, and this may indicate that certain vehicles don’t reliably report their departure from Dufferin Loop. The data gives the impression that they leave much earlier than they actually do. By King Street, however, everything is back in sequence.
What is quite visible on this chart is the same sort of bunching that we see on other routes with vehicles running in pairs. This is particularly evident in the evening causing wide gaps in service.
The headway charts show the same behaviour we saw earlier on the King and Queen lines. Buses leave Dufferin Loop (first page of the northbound charts) on widely varying headways even when the average (and scheduled) headway is under 4 minutes. The plus-or-minus three minute rule clearly does not apply here and headways over 9 minutes are common. Many headways are at one minute or less showing that buses leave the terminal together.
The timepoints I have chosen for Dufferin were selected to avoid areas where it is difficult to resolve the location of the buses due to correction effects. The timepoints are:
- Dufferin Loop
- King Street (useful for calculating round trip times to Dufferin Loop and, hence, layover times)
- Dundas Street (south of the congested area around Dufferin Mall)
- Dupont Street (north of the subway connection and the end of a congested area on the route)
- Eglinton Avenue
- South of Lawrence (chosen because data at Lawrence itself is unreliable)
- Wilson Avenue & Dufferin Street
- Wilson Station
The irregular headways persist over the length of the route, although they do not open up into wider gaps as we move from one end of the line to the other as they do on King. Note that headways in the evening when the schedule is 8 to 9 minutes continue to be erratic, but the swings (even those within the TTC’s three-minute target) produce wide gaps in service.
Link times are consistent over the route as one would expect on a holiday. The spikes in the northbound (up) trips from Dufferin to King are caused by the problem mentioned above where the departure time from the loop is not reported, and the last known time for a bus at the loop is much earlier than it really was. A similar problem happens southbound from Wilson Station.
In the next installment, I will look at overall headway and link times for the month to see the general pattern of operations before turning to a few specific days as examples.