In comments on a recent article, a reader asked me to review the 100 Flemingdon Park bus route. As it happens, I have been tracking the route for almost a year and so have lots of data.

The picture is not a pretty one. This route has been affected by construction at various times, but the major problem is with headway management, or rather a lack of any, on some days. Very large gaps with bunches of three or more buses are common.

If nothing else, the Flemingdon Park shows that you don’t need streetcars fighting downtown traffic to have chaotic service. Too often there are assumptions about “typical” problems without looking too hard for the details.

Service on this route is unreliable during most of the 11 months covered by the data here. Erratic headways also generate uneven loads on buses. The theoretical capacity of a route might be capable of providing comfortable service, but with bunching most riders have longer than scheduled waits and crowding that is worse than the theoretical average. Riders complain about actual service conditions. Management responds with averages that hide the problems while claiming that they meet the “board approved” service standards most of the time.

Scheduled Service

The schedules in effect on June 1, 2023, saw the Flemingdon Park buses running between Don Mills & Eglinton south to Broadview Station. A former link west to Eglinton Station was replaced by a separate 34C Eglinton East bus to isolate the 100 service from construction delays on Eglinton.

On June 18, 2023, the southern terminus shifted from Broadview Station to Pape Station and the route was interlined with 72A Pape. This change was part of a general restructuring of routes during construction at Broadview Station and the surrounding neighbourhood.

On November 19, 2023, the Flemingdon Park buses returned to Broadview Station, and this has been their terminus ever since. Although a new schedule was issued in May 2024, the service details are the same except that the route now operates from Eglinton rather than from Birchmount garage.

The remainder of this article looks at various months over the June 2023 to May 2024 period in detail to see what the service actually looked like from a rider’s point of view.

Overall Headway Reliability

Headways on 100 Flemingdon Park have been unreliable over the entire period. The charts below show the median, 85th percentile and maximum headways for all weekdays.

• For the north end, the screen line is just south of Eglinton southbound. Due to Line 5 construction, the route around the loop at Don Mills varied from time to time. I have use a screenline on the stable part of the route.
• For the south end, the screenline is at Broadview & Pretoria, one block north of Broadview Station Loop, for the period when the route operates to that station.
• For the period when routes 100 and 72 interline, the screenline is at Lipton Ave. north of Pape Station.

These locations ensure that departures are measured clear of the loops where layovers and other bus shuffling might occur.

The median values (blue) lie generally close to the scheduled service level. This is what one would expect with well-behaved service. However, in some cases, the median can drop below the average. For example, if service is badly bunched, then there are a few very high maxima (big gaps) but a lot of very short headways (buses bunched together). The average would stay roughly at the scheduled headway, but the much more common bunched vehicles will dominate the stats and lower the median value.

The contrast between the TTC’s official service standard, based on averages over an hour, and actual rider experience is illustrated by the blue medians and the spread between them and the higher percentiles.

The maximum values (red) show the highest headway value observed within an hour on one day. When I first set up these charts, it was surprising that I had to use a y-maximum value of 45 minutes to fit in most of the spikes. Maxima over 20 minutes are common, and over 30 minutes often enough that they cannot be dismissed as an occasional issue due to special circumstances.

The orange line is the 85th percentile showing the level below which most headways lie. Ideally this should be close to the median indicating that the underlying values are closely grouped.

Note how the lines spread out and become more chaotic as the day goes on. The 8am and 1pm data are not too bad, but by the PM peak and into the evening, the 85th percentile and maximum values pull further away from the median. There is some recovery by late evening, but the service is by no means ideal.

There is some improvement in the evening period after the service returns to Broadview Station in November 2023, but this benefit is undone in Spring 2024 due to a combination of construction enroute and an absence of line management. Later in the article, I will review some individual days’ operation to show how badly bunched the service can be.

The June 2023 Transition from Broadview to Pape Station

On June 18, 2023, the southern terminus of the route shifted east from Broadview to Pape Station, and the route was interlined with 72 Pape. At the north end of the route, the loop from Wynford was revised due to Ontario Line construction in the week of June. Headway reliability declined as charts in this section show.

The chart below shows the overall monthly statistics for service northbound from Danforth. Weeks 1-3 are from Broadview Station, and weeks 4-5 are from Pape Station.

Although the average headways are the same for all five weeks, there are differences in the standard deviations with weeks 1-3 being lower than the others.

Standard Deviation values of five or more show that over half of the service operates on headways over a wide range, typically in a band 2 SD wide. The remaining service lies outside this band. Near the terminus, one would expect to see an SD of no more than a few minutes.

At a more detailed level, the following charts show week-by-week behaviour.

• On the left is the raw data with each dot representing the headway of one vehicle. These are colour coded by day, and there is a trend line showing the overall lie of each day’s data.
• On the right are block-and-whisker charts showing the quartiles of headway values. The boundary between the green and blue boxes is the median value. Half of all trips are included in those boxes (the 2nd and 3rd quartiles), while the lower red tails hold the 1st quartile and the upper mauve tails hold the 4th quartile. Note how long the 4th quartile tails are later in the month and there are maxima above half an hour.

Headways on the 100 Flemingdon Park leaving the subway northbound were spread over a very wide range of values in June 2023 from week 4 onward when the 72/100 route was interlined. In weeks 1-3, headways are less scattered. although there is a fair amount of bunching in the PM peak reflected by data points near zero.

The increased range and scatter of headways are typical of the middle of a route, but this shows how even when buses come to an off-street terminal (Pape Station in this case), buses are not dispatched to restore even headways.

Weekend service does not fare any better. The trend lines up to June 18 lie higher on the charts than later dates because the combined 72/100 scheduled service was generally more frequent than the separate 100 route. However, the scatter of headway values even when the 100 Flemingdon Park buses ran on their own is appalling with many very high and very low values characteristic of bunched service.

Service From Pape Station September 2023

September 2023 has an entire month of operation from Pape Station. The standard deviation values are high as is normally seen on a mid-route location where buses manage to catch up with each other. The 72 Pape bus is somewhat erratic and this carried over into the interlined 100 Flemingdon Park trips. That high SD value typically indicates a lot of bunching with wide gaps.

The service is unreliable, but TTC metrics only measure at the “terminal” which for the combined 72/100 service is the bottom end of the 72 Pape bus route.

The week-by-week data show a wide range of headways all the way from bunched service to gaps of 15 minutes and beyond.

Weekends show the same pattern with many very short headways (bunching) and wide gaps.

The November 2023 Transition from Pape Station to Broadview Station

For Weeks 1-3 of November 2023, 100 Flemingdon Park operated via Pape and interlined with 72 Pape at the subway station. From Week 4 onward, 100 Flemingdon Park operated on its normal route to Broadview Station after construction there completed. The scheduled service was less frequent in some periods starting in week 4, but the standard deviations in headway values were also better showing some improvement in reliability. This pattern is the reverse of the one seen in June 2023 when the routes were combined.

At a detailed level, wide headways were more common northbound from Pape Station in Weeks 1 to 3 as shown both by the scatter in the plots on the left and the wider quartiles in the charts on the right.

[Note: In week 4, there was no tracking data on Friday, November 24 until about 9:30 am. This causes the trend line to “swoop in” from above.]

As with the June 2023 transition, the period of separate 100 Flemingdon Park operation shows higher weekend headway trend lines with the return to the wider scheduled headways. However, the scatter in data points remains high even as a separate route.

March 2024

March 2024 was a comparatively stable month. Service operated from the regular southern terminus at Broadview Station and the line was not subject to construction delays. The standard deviation in headways at most times was low, only a few minutes, except during the early part of the PM peak.

Looking at the weekly and daily breakdowns, the lower SD values are reflected in data points for individual headways that are clustered more tightly around trend lines, and the relatively small range of the 2nd & 3rd quartile values (the central bars in charts on the right) compared to other months.

Weekends in March were not quite so well-behaved, and headways well above the scheduled level of 9-10 minutes are fairly common.

Wednesday, March 20

This is an example of a fairly well-behaved weekday in mid-March. Each line represents the movement of one bus back and forth on the route. The slope of the line reflects the vehicle speed, and there are few examples of chronic congestion where the lines slope more horizontally at problem locations. (For example near Linkwood and at Don Mills/Overlea in thePM peak.

Buses get a layover at one or both ends of their trips.

Bus spacing stays fairly regular until after 1pm when some pairing of buses can be seen. Even when this occurs, most runs stay spaced apart, and the pairs do not persist over several trips.

(Note: The loop at Concord Place is mapped as a single location and so buses appear to sit at one place while making their trips north and south there.)

Saturday, March 9

Service on March 9 stays quite regular until 10am when problems begin including:

• A persistent gap caused by one bus running late as a pair with ts follower (dark blue and turquoise lines from about 10:30 southbound from Don Mills. This gap persists until about 12:20 when “dark blue” short turns at Concorde Place to take its proper place.
• Bus pairing continues through the afternoon and by 4pm there are two pairs running together.
• A triplet of buses travels as a pack from about 5pm to just after 7pm.
• Layovers at terminals are short in the afternoon period implying a too-tight schedule, especially with little sign of congestion.

By early evening, service is more regularly spaced although there is still some bunching.

May 2024

In May, the SD values have crept up compared to March, particularly from week 3 onward when they overlap the averages indicating severe bunching. This corresponds to construction to reconfigure lanes between O’Connor and the Leaside Bridge. However, even in weeks 1 and 2, the headway values are more scattered than in March.

Weekly breakdowns illustrate the wider scatter of headway values especially from week 3 onward.

Weekends, especially Sundays, saw many very wide gaps and much bunching. Detailed charts for two of these days are later in the article.

Thursday, May 16, 2024

The effect of construction north of O’Connor is evident during the PM peak as the slope in the lines changes abruptly from about 3:30pm to just after 6pm. The effect is stronger northbound than southbound.

However, this does not explain the condition of service earlier in the day when buses travel in groups of two or more. In some cases, buses arrive and leave from terminals together even though they could have been spaced apart. This problem reaches ts height from 3:30pm onward where the combination of bunching and the construction delay produces very wide gaps in service.

The worst of these is between 4:20 and 5:20pm when no buses leave Broadview Station. This is entirely due to the “no short turn” approach to line management that could otherwise have broken up packs of buses to restore more reliable headways. Some service spacing is evident northbound after 5pm, but the effect of that wide gap persists until after 8pm.

Friday, May 24, 2024

May 24 shows a similar pattern to May 16 with bunching and wide gaps opening up in the PM peak. Buses run in packs, and this persists until about 10pm.

Saturday, May 4, 2024

May 4 shows a problem we saw earlier with short layovers at terminals in the afternoon period indicating tight schedules. However, there is only minor traffic congestion south of Millwood to complicate this problem.

Service begins to bunch mid-morning, but the situation becomes dire through the afternoon and early evening causing wide gaps in service until after 8pm. There is some short turning and express operation, but this does not keep up with the degree of bunching.

Sunday, May 26, 2024

May 26 shows a similar pattern to May 4 above with buses running in packs of two or more from mid-morning to late evening. Some short turns occur, but they do not always break up groups of buses and occasionally contribute to packs running in the opposite direction (see the period from 2-4pm in Thorncliffe Park).