US & Canada light rail & streetcar systems ranked by riders/mile

[mhays]

For starters, transporting 10 people vs 1 might mean 9x more cars off the road.

[Nouvellecosse]

I'm not sure I see the real advantage in preventing 10 cars from driving 1km as opposed to 1 car driving 10km. Sure, you could make the argument that longer distance commuters will spend more of the trip on the highway and get better fuel economy, but people living in dense urban settings tend to be less likely to drive cars anyway, so there's less chance you'd even be replacing ten 1km car trips.

[electricron]

Originally Posted by Nouvellecosse

I'm not sure I see the real advantage in preventing 10 cars from driving 1km as opposed to 1 car driving 10km. Sure, you could make the argument that longer distance commuters will spend more of the trip on the highway and get better fuel economy, but people living in dense urban settings tend to be less likely to drive cars anyway, so there's less chance you'd even be replacing ten 1km car trips.

I don't think it really matters, a passenger kilometer can measure how much pollution is being saved. Whether its one passenger over 10 kilometers, or ten passengers over 1 kilometer, it's still 10 passenger kilometers worth of pollution saving. Transit can be used to reduce highway and parking congestion, and it can be used to reduce pollution. Number of passengers may be the better statistic to use for traffic congestion, passenger kilometer may be the better statistic to use for pollution avoidance.
They're both great statistics that should be monitored if possible. On systems that have accounting processes to count the total passengers boarding and alighting, it's difficult to measure how far they ride the transit vehicle. On systems that have processes that account for when each passenger boards and alights the transit vehicle, it is possible to measure how far they ride the transit vehicle. Not every transit system can give a correct passenger kilometer statistic, which is why it isn't always used or reported.

[Nouvellecosse]

I don't think it's necessary to find tool that can precisely measure the exact trips; for most intents and purposes you can get similar accuracy simply using surveys and market research.

[eixample]

The Philadelphia total light rail system length of 68.4 miles is off. I realize these numbers come from wikipedia's light rail ridership page, but wikipedia also lists the three components of SEPTA's light rail separately as 19.8 miles for the five subway surface routes, 8.4 miles for the route 15 trolley and 11.9 miles for the two suburban trolleys. That would bring the total length at 40.1 miles and riders per mile at 2,783. Even if you counted the shared tunnel for the five subway surface lines for each subway surface line individually (the green line and muni metro are similarly laid out so you'd want to measure it the same way as they do it), you still would only get to 51.3 total miles or 2,175 riders per mile.

Regardless, it is worth noting that the two suburban lines have lower numbers so that drags the numbers down. The route 15 has decent ridership but not as good as the subway-surface routes that run frequently and are often very crowded (per SEPTA's 2015 numbers, the five subway surface routes have ridership in the 15,000 to 18,000 riders per week day range, which translates to around 2,500 to 2,600 riders for mile except for the shortest of these routes, route 34, which has ridership around 3,400 riders per mile).

[shivtim]

Interesting, thanks for compiling. But I think there are some inconsistent data that really affects the numbers. For example, the Atlanta streetcar is listed at 2.7 miles long. That's the total length of all of the track, which operates one-way on paired streets (like parts of Portland's system). But then the DC streetcar is listed at 1.9 miles, while on wiki it's listed as both 2.2 and 2.4 miles, but really the total track length is double that number. Same story for systems like SLC's S line. It's listed as 2 miles, but that's really 4 miles of track.

If something operates more as a loop the total track length is counted, but if something is two-way on a linear path only half the actual track length counts. This makes ridership appear to be 50% of what it actually is on one-way paired systems.

My point is, if this is intended to be a useful comparison across systems, the denominator has to be the same. What does "per mile" mean? Is it per track mile, or per route mile?

Also a quick correction on the ridership for the Atlanta streetcar, since it's not from the wiki source. It's from here. He says "about 1000 per day," which is true, but the exact number is 1105 per day (100,608 passengers in the first 91 days of 2016). Also of note is this is daily ridership, not weekday ridership, but as far as I know the actual daily counts have never been released. One final note is that 2015 ridership was much higher, when the system was free.

Just to show how big of a difference all of these assumptions and data choices make:
-Using the data in your spreadsheet now, the Atlanta streetcar is listed at 370 daily riders per mile.
-Using a route length of 1.35 miles and ridership numbers from 2015, it would be 1641 daily riders per mile.

[Bikemike]

Originally Posted by Cirrus

You're right. I copypasted from Wikipedia, but have corrected them now.

Eh, it tells you something different, but it rewards long sprawling commuter trips and punishes shorter urban trips. It tells you something different, but I wouldn't say better.

Well, it tells you there is also an apples to oranges comparison going on. Your comment fails to consider the fact that longer rail lines aren't also without shorter segments which contain a large number of shorter urban trips (high ridership density segments). The difference is some systems just aren't built as far into the suburbs and therefore tout higher averages, which can be misleading when compared with lines that run through highly urbanized areas as well as extensive suburban stretches.

In the end, when comparing apples and oranges, neither can be called "better"

[Cirrus]

Originally Posted by shivtim

But I think there are some inconsistent data that really affects the numbers. For example, the Atlanta streetcar is listed at 2.7 miles long. That's the total length of all of the track, which operates one-way on paired streets

That's absolutely an inconsistency. A quick wikipedia study is going to produce problems like that. See any others?

Originally Posted by Bikemike

Your comment fails to consider the fact that longer rail lines aren't also without shorter segments which contain a large number of shorter urban trips (high ridership density segments). The difference is some systems just aren't built as far into the suburbs and therefore tout higher averages, which can be misleading when compared with lines that run through highly urbanized areas as well as extensive suburban stretches.

It's not "misleading" in the slightest. The entire point of doing it is to normalize an inherently apples-to-oranges comparison and make it as apples-to-apples as mathematically possible. One of the main points of this type of comparison is so you can see how lines with different characteristics produce higher or lower ridership.

[Chef]

Minneapolis is helped by the fact that the Green Line between the downtowns (which is 11 miles) goes almost entirely through urban areas, and nearly all of the Blue Line is within Minneapolis proper except for stops at the airport, the Mall of America and a Bloomington office park. The two lines do manage to connect a lot of destinations - in addition to the Mall, the airport and the downtowns, they also hit the U of M and all the major professional and college sports stadiums.

[shivtim]

Originally Posted by Cirrus

That's absolutely an inconsistency. A quick wikipedia study is going to produce problems like that. See any others?

Yep, and again thanks for putting this together, it's interesting. I'm not familiar enough with all of the systems to notice issues, the only ones I saw were the length and ridership of the Atlanta streetcar, and the length of the DC streetcar.

I'd be interested to see how some airport terminal lines and other systems compare. For example, the ATL skytrain (which is technically a BRT), the Detroit people mover, and the Disney World monorail.