This article covers service on the R2 over a period of four weeks from September 30th to October 27th. The data was retrieved from Transsee. I was heavily inspired by Toronto's Steve Munro to start this series, and hopefully shed some light on OC Transpo's hideous reliability.
Cancellations
- Sept 30: 1 SB, 1 NB
- Oct 1: 4 SB, 5 NB
- Oct 2: 4 SB, 4 NB
- Oct 3: 7 SB, 7 NB
- Oct 4: bad data
- Oct 5: 1 SB, 1 NB
- Oct 6: 0 SB, 4 NB
- Oct 7: 5 SB, 6 NB
- Oct 8: 9 SB, 8 NB
- Oct 9: 4 SB, 4 NB
- Oct 10: 3 SB, 0 NB
- Oct 11: 12 SB, 9 NB
- Oct 12: 0 SB, 0 NB
- Oct 13: 0 SB, 0 NB
- Oct 14: 0 SB, 1 NB
- Oct 15: 5 SB, 6 NB
- Oct 16: 11 SB, 9 NB
- Oct 17: bad data
- Oct 18: 0 SB, 0 NB
- Oct 19: 15 SB, 15 NB
- Oct 20: 8 SB, 9 NB
- Oct 21: 15 SB, 12 NB
- Oct 22: 11 SB, 8 NB
- Oct 23: 0 SB, 0 NB
- Oct 24: bad data
- Oct 25: 0 SB, 0 NB
- Oct 26: 4 SB, 4 NB
- Oct 27: 3 SB, 4 NB
This route runs:
- SB Weekday: 99 trips
- SB Saturday: 72 trips
- SB Sunday: 65 trips
- NB Weekday: 96 trips
- NB Saturday: 73 trips
- NB Sunday: 64 trips
The cancellation rate during this period is 5.5% southbound and 5.4% northbound in this period, which is quite terrible. However, this hides the variability, in which some days have no cancellations, and other days may have 15% cancellations in certain directions, or even more.
The goal should be to reduce these to 0.5%, or 0%, if that's feasible. However, after City Council had OC Transpo dispose of over a hundred buses, they - surprise! - don't have enough buses to maintain scheduled service.
Schedule Deviation
These charts show the average headway for each hour over the period of a working week (does not include weekends), as well as the deviation from each headway. OC Transpo's "service delivered" masks service which is functionally useless to the rider because it shows up in pairs or trios. The higher the dotted line, the more off-schedule the route is at that point in time.
Week 1 is Sept 30-Oct 4; week 2 is Oct 7-Oct 11; week 3 is Oct 14-Oct 18, and week 4 is Oct 21-Oct 25.
Northbound, service usually starts to get off schedule in the early morning, followed by a midday lull, and then returning with more unreliability in the afternoon peak and into the evenings. The range of deviation is quite high, indicating different on-time performances for each week of this analysis.
Southbound, the route generally starts with even service, but the schedule deviation rises until it reaches nearly the scheduled service by the afternoon peak, resulting in often-bunched service. Service in week 4, October 21st to October 25th, also has irregular service in the early morning and the evenings.
On-Time Performance
These charts shows on-time performance over the five days of Monday, September 30th to Friday, October 4th. However, these charts, which show abysmal performance already, hide variations in schedule uncertainty. Please note that this only encompasses one week; I'm using it as a proxy for overall travel time.
To see a better picture of where delays occur, we need to look at variation in time.
These charts show that AM Peak performance is okay, if not great. Northbound, schedules need to be adjusted between Heron and Carling, which show a large source of delay. Similarly, southbound seems to gain time between Walkley and Greenboro; delays appear between Carling and Carleton, and then somewhat abate south of Heron.
In the PM peak, delays are far more serious; over the week, Transsee shows the entire service delayed from Carleton to Carling going northbound; southbound trips run 10 minutes delayed or more, half the time. Half the time. This delay does not occur at the beginning of the route - OC Transpo needs to allocate a longer runtime between Gladstone and Carleton, where the delay appears.
Let's look at string diagrams for one day.
String Diagrams: Wednesday, October 2nd
This is only one day, but this route specifically is free for all users; you can check any day you want, and see what appears for that day.
I chose a Wednesday, as it's the busiest day of the week; some days are better behaved, such as weekends, but this is instructive to see how the service behaves to riders on the bus (or not on the bus).
We can see that the service starts well spaced in the morning, but at the end of the AM Peak, the service begins to bunch by 9:00 AM, and remains bunched until the the late evening, when the trips get better spaced again. This could be indicative of problems on the route, but I believe this is a problem stemming from bad schedules on interlined routes; very few blocks run into others on this route, as trips generally start bunched and stay that way through their trip.
The chart also shows where services get delayed, and can be sped up. Steeper lines indicate higher speed, and conversely, shallow lines show a slow moving bus. We can see problem spots on the chart:
- Gladstone is a problem spot in the peak periods going southbound.
- Many southbound trips, all day, get delayed making the left turn from Preston onto Carling.
- Dwell times at Carleton are a problem in both directions.
- Southbound trips are delayed trying to turn left across Heron twice; first, from the Bronson off-ramp to Heron, then across Heron onto Data Centre.
While I did not analyse any other string diagrams, a casual glance at other days reveals similar patterns of bunching and delays.
This is it for my first route analysis article. Please leave any comments or suggestions in the comments, and I'll hopefully be back soon with another article.
No comments:
Post a Comment