[afransen]

I'm quite familiar with the Erb/Needles roundabout. Which in my mind is very poorly designed, with the approaches not on enough of an angle, the pavement markings poor, and the pedestrian crossings not far enough from the roundabout. I don't know why the paint markings for the lanes, don't extrude out onto the roundabout. Or the bullheads aren't more obviously coloured.

I drove around that roundabout backwards once, as a passenger, not long after it opened; they were trying to turn left (west to south). Which should be nigh impossible if the thing was properly designed. Looking at it in Streetview, I'm surprised nothing has changed, other than making the pedestrian crossing signs larger.

The two-lane roundabouts do seem more problematic. And if that (quite old) roundabout still meets MTO design standards, then MTO design standards need to be updated.

That simply may not be an appropriate place for a 2-lane roundabout. Frequently such things have in relatively busy pedestrian locatons have pedestrian signals.

Are there any MTO roundabout in similar environments, with shops on each corner, and very close residential? If you look in the UK, for example, such locations either seem to have pedestrian-activated lights.

Is there a link to that? I'm familiar with KPIs, but not KSIs. There's no roundabout data there right?

I've used this intersection in the last few months. It was pretty backed up. I would agree it is not well-designed. Looks like a good candidate for a turbo roundabout. The pedestrian crossings maybe should be set back further with better refuges. Not sure about the average volumes, but this intersection may warrant a signalized intersection.

 

[txlseries4]

Is there a link to that? I'm familiar with KPIs, but not KSIs. There's no roundabout data there right?

The data I'm grabbing is from here, but I'm querying it after having put it into a database: https://data.torontopolice.on.ca/datasets/TorontoPS::pedestrian-ksi/about
It can be downloaded as a .CSV and opened up as a spreadsheet, but refining the data might not be straight-forward.

TPS has a Power BI dashboard here with limited information: https://data.torontopolice.on.ca/pages/pedestrians
Some of the numbers are different on the dashboard because they only include injuries which are major or fatal.

There's also a Vision Zero dashboard here: https://www.toronto.ca/services-pay...zero-dashboard/seriously-injured-vision-zero/

There's nothing that directly mentions roundabouts in the data that I've seen, but from a quick google search found there's one at Coe Hill and Windermere in Etobicoke, and another (small) one at Yeomans Rd and McAllister Rd near Sheppard and Bathurst, so I looked for collisions at those spots. No collisions at the one in Etobicoke, 6 collisions at Yoemans and McAllister between 2014-Sept 2024, only 1 involved a pedestrian but didn't involve an injury. But both are nestled in neighbourhoods so not a big surprise. Hardly like the awful ones in KW which I've driven on too

 

[nfitz]

The data I'm grabbing is from here, but I'm querying it after having put it into a database: https://data.torontopolice.on.ca/datasets/TorontoPS::pedestrian-ksi/about
It can be downloaded as a .CSV and opened up as a spreadsheet, but refining the data might not be straight-forward.

TPS has a Power BI dashboard here with limited information: https://data.torontopolice.on.ca/pages/pedestrians
Some of the numbers are different on the dashboard because they only include injuries which are major or fatal.

There's also a Vision Zero dashboard here: https://www.toronto.ca/services-pay...zero-dashboard/seriously-injured-vision-zero/

There's nothing that directly mentions roundabouts in the data that I've seen, but from a quick google search found there's one at Coe Hill and Windermere in Etobicoke, and another (small) one at Yeomans Rd and McAllister Rd near Sheppard and Bathurst, so I looked for collisions at those spots. No collisions at the one in Etobicoke, 6 collisions at Yoemans and McAllister between 2014-Sept 2024, only 1 involved a pedestrian but didn't involve an injury. But both are nestled in neighbourhoods so not a big surprise. Hardly like the awful ones in KW which I've driven on too

I don't think local data would tell us much.

I didn't know about the Etobicoke roundabout - which is quite small. That one on Yeomans and McCallister isn't a roundabout. It has stop signs at the intersection. You see these all the time in some cities, but I've never heard them called a roundabout before. Traffic circle occasionally - but very very tiny.

 

[txlseries4]

I don't think local data would tell us much.

Yea, I agree. I've looked before for provincial and national data and it just doesn't seem to exist in a way that's easy to consume (vis-a-vis importing data into a database, or dashboards).

That one on Yeomans and McCallister isn't a roundabout. It has stop signs at the intersection. You see these all the time in some cities, but I've never heard them called a roundabout before. Traffic circle occasionally - but very very tiny.

ha, I missed all 4 stop signs on google maps, appreciate pointing that out

 

[BrenWilson]

Not sure about the average volumes, but this intersection may warrant a signalized intersection

Yes. The issue with roundabouts in Ontario is multifaceted. We design them poorly, we put them in the wrong places, and our drivers don't act the same as european drivers.


I'll use the dutch as the traffic design golden boy for this. Roundabouts there are more often than not used on suburban roads with only one lane per direction, they are radial and not tangental, and have well set back crossings where its reasonable to expect drivers to yield. Two lane turbo roundabouts there are rarer and typically feature completely grade seperated pedestrian/cycle infrastructure. Generally however they still design multi lane junctions as signalized intersections.


In Ontario we almost never have suburban roads (not streets) with only one lane per direction. Generally we start with 2 lanes per direction on roads, which already conflicts with the dutch examples. We really only have one lane per direction roads in rural settings or local streets. You also rarely see radial roundabouts (though the MTO guidebook does recommend them, TAC design guides don't really touch on them); basically every roundabout here is tangential which encourages not stopping. We then add chicanes to slow drivers leading up to the circle but this quite frankly just complicates the entire thing and doesn't likely help pedestrians.



So can roundabouts be safe for pedestrians? Sure, theoretically. We will ever see them here? Unlikely without significant changes to our entire road design methods.

 

[afransen]

Yes. The issue with roundabouts in Ontario is multifaceted. We design them poorly, we put them in the wrong places, and our drivers don't act the same as european drivers.


I'll use the dutch as the traffic design golden boy for this. Roundabouts there are more often than not used on suburban roads with only one lane per direction, they are radial and not tangental, and have well set back crossings where its reasonable to expect drivers to yield. Two lane turbo roundabouts there are rarer and typically feature completely grade seperated pedestrian/cycle infrastructure. Generally however they still design multi lane junctions as signalized intersections.


In Ontario we almost never have suburban roads (not streets) with only one lane per direction. Generally we start with 2 lanes per direction on roads, which already conflicts with the dutch examples. We really only have one lane per direction roads in rural settings or local streets. You also rarely see radial roundabouts (though the MTO guidebook does recommend them, TAC design guides don't really touch on them); basically every roundabout here is tangential which encourages not stopping. We then add chicanes to slow drivers leading up to the circle but this quite frankly just complicates the entire thing and doesn't likely help pedestrians.



So can roundabouts be safe for pedestrians? Sure, theoretically. We will ever see them here? Unlikely without significant changes to our entire road design methods.

The current government is proving to be unhelpful in this regard. Many suburban roads (collectors, not arterials) could probably be just one lane per direction and not use signalized intersections. The two lanes per direction is mainly to provide dedicated turning lanes and queuing space, which shouldn't be necessary with roundabouts. Lots of room for 4 to 3 or 4 to 2 road diets, ideally with space given to bike lanes, if we were to use roundabouts for such roads. The collector street I live off of is one lane per direction with bike lanes, but the cross streets mostly are 2 lanes per direction and signalized. I think all the signalized intersections could be swapped with roundabouts.

 

[BrenWilson]

The current government is proving to be unhelpful in this regard.

Oh that goes much deeper than the current government. A progressive (in transportation design) government could certainly change this to some degree but its just how industry "best practise" works here. You basically need to rewrite national, provincial and local design guides from the ground up to actually see permanent change in that. The overwhelming view is that roads are for cars, and bigger/more is better; window dressing is paid to safe design and pedestrian/cycling facilities but its a cultural issue in the profession and the only progress you see are with hard fought constant battles at the municipal level.

 

[Monarch Butterfly]

In Europe, pedestrian crossings at intersections tend to be setback from the corner. At least a car length. This allows motorists to make a turn and stop before the crossing, but still be able to clear the cross-traffic. Crossings are not at the corners, but away, which allows long trailers turning right to avoid pedestrians.

 

[reaperexpress]

Oh that goes much deeper than the current government. A progressive (in transportation design) government could certainly change this to some degree but its just how industry "best practise" works here. You basically need to rewrite national, provincial and local design guides from the ground up to actually see permanent change in that. The overwhelming view is that roads are for cars, and bigger/more is better; window dressing is paid to safe design and pedestrian/cycling facilities but its a cultural issue in the profession and the only progress you see are with hard fought constant battles at the municipal level.

There is talk about changing our current practices. For example this talk at the upcoming Transportation Association of Canada conference:

Radial single-lane design is pretty much a prerequisite of bicycle priority roundabouts so that will presumably come up in this discussion.

 

[Monarch Butterfly]

From https://www.facebook.com/share/p/16UPW8uVrG/

 

[BrenWilson]

Radial single-lane design is pretty much a prerequisite of bicycle priority roundabouts so that will presumably come up in this discussion

I hope so, as depsite being listed as fair game in OTM books, there is no radial design guide in the TAC manual and basically no one in ontario uses them - everything is tangential. I find this amusing as Ontario tangential roundabouts often feature elaborate chicanes in an attempt to slow approaching traffic which in my mind places less of the driver's attention on potential pedestrians, making the design more complex and less safe.

 

[nfitz]

View attachment 659855
From https://www.facebook.com/share/p/16UPW8uVrG/

And what about pedestrian deaths?

There's not even any sidewalk in that picture!

 

[txlseries4]

And what about pedestrian deaths?

There's not even any sidewalk in that picture!

What do you mean, it's right there in the middle. Pedestrians aren't allowed to leave the inner circle though. Sorry!

I think the photo was taken on a road that doesn't have pedestrians, like a country road coming into the city, but it might actually be outdated; I tried looking on Google Maps but couldn't find any spot where that would be because Carmel's now surrounded by other towns and sprawl without big fields between. All the roundabouts I saw on Maps also accommodate pedestrians and look like these:

 

[BrenWilson]

Carmel is probably not the best example to look at vs Toronto given what Carmel is (a small satellite city/suburb); but I commend them for actually throwing off the status quo when it comes to road design - wholesale and not just a few select roads.

 

[reaperexpress]

There's nothing that directly mentions roundabouts in the data that I've seen, but from a quick google search found there's one at Coe Hill and Windermere in Etobicoke, and another (small) one at Yeomans Rd and McAllister Rd near Sheppard and Bathurst, so I looked for collisions at those spots. No collisions at the one in Etobicoke, 6 collisions at Yoemans and McAllister between 2014-Sept 2024, only 1 involved a pedestrian but didn't involve an injury. But both are nestled in neighbourhoods so not a big surprise. Hardly like the awful ones in KW which I've driven on too

As far as I'm aware, here are all the roundabouts in the City of Toronto.
I define a roundabout is designed as a circulating intersection with yield or stop on all approaches, and no yield/stop for traffic in the circle. I haven't included the neighbourhood traffic calming circles because the traffic control seems unrelated to the presence of a circle - they act as all-way stop or two-way stops rather than an actual roundabout.

Lowe's Place & Resources Rd:

Gilmour Ave & Maher Ave:

Windermere Ave & Coe Hill Dr & Budgell Terrace & Olaves Rd

Connaught Circle:

Roxborough Dr & Highland Ave:

All of these are single-lane roundabouts on low-volume roads and some of them even have a low-speed radial layout. So I wouldn't expect to find any KSIs in the database.

The Roxborough roundabout has way too much asphalt, here's a concept for an adjustment to create a slower and more intuitive layout: