Cartogram of Amtrak Ridership by State, 2022

[Matthew H Fish]

The other day, I posted an old cartogram I made in another thread. I started thinking about cartograms and wondered what a cartogram of Amtrak's ridership would look like. So, using the 2022 figures, I made this:

I know the aesthetics and organization are a little confusing---but basically I didn't want to polish it before knowing if it was a good idea conceptually.
In this diagram, each square represents 100,000 Amtrak riders. If a state had Amtrak service but less than 100,000 riders (which many did), I still gave them a square. South Dakota and Wyoming are left off.
I tried to group the states together and keep geographical sense, but as is often the case with cartograms, that becomes difficult to do when the difference from geographical size to conceptual size becomes so great.
While an easy take-away from this map is that the NEC corridor is very important in Amtrak ridership, another point is that corridor routes in general are important. Of states without corridor service, only Florida, Texas and Colorado get more than a single square. Even states with less intensive corridor service, like Washington/Oregon or Illinois to Wisconsin/Michigan/Missouri, still are much bigger than long distance routes through big cities.
Any suggestions for how to improve this map?

 

[Trollopian]

Very interesting, Matthew. (From someone who overwhelmingly rides on the NEC and the Cap, but loves seeing other landscapes. "Flyover country," what an insult.) What does ridership represent in your graph: origin or destination? Presumably not both, to avoid double-counting.

And as you've pointed out elsewhere (in the thread on "Frontier and Remote [FAR]," among others), the fact that a state generates little ridership doesn't mean that Amtrak is unimportant there. In fact quite often the reverse.

 

[Matthew H Fish]

Very interesting, Matthew. (From someone who overwhelmingly rides on the NEC and the Cap, but loves seeing other landscapes. "Flyover country," what an insult.) What does ridership represent in your graph: origin or destination? Presumably not both, to avoid double-counting.

And as you've pointed out elsewhere (in the thread on "Frontier and Remote [FAR]," among others), the fact that a state generates little ridership doesn't mean that Amtrak is unimportant there. In fact quite often the reverse.

I should actually know what that means----but I don't.
I took it from the State Fact Sheets, like this one:
https://www.amtrak.com/content/dam/...ments/corporate/statefactsheets/NEWYORK22.pdfBut now that I look at it, it doesn't describe what those ridership numbers mean.

And yes, ridership doesn't reflect necessarily how vital a ride is. This metric also doesn't look at passenger miles, which would probably equalize the figures at least a little bit.

 

[jis]

I should actually know what that means----but I don't.
I took it from the State Fact Sheets, like this one:
https://www.amtrak.com/content/dam/...ments/corporate/statefactsheets/NEWYORK22.pdfBut now that I look at it, it doesn't describe what those ridership numbers mean.

And yes, ridership doesn't reflect necessarily how vital a ride is. This metric also doesn't look at passenger miles, which would probably equalize the figures at least a little bit.

I believe it is simply the sum of segments that originate and terminate at the station. Connected trips are counnted in terms of individual segments, So Tampa gets all the Thruway transfers on its account as arriving/departing Tampa by train. I don;t think the counting is any more sophisticated than that.

 

[Septa9739]

State fact sheets report unlinked trips as JIS described, so as much as 10% of traffic at certain major connecting stations gets double counted. Think CHI, PGH, NYP. That certainly would pull down some of the inflation in the “big train states.” Unfortunately Amtrak does not release information on linked trips leading bored weirdos like myself to make guesses of scraps of information that do get released.

To vitality, I think a really interesting visual, maybe expressed as a cartogram, would look at the ratio of riders to a station city’s population. That paints a very different story. New York is about 1, Philly 2, DC 5, Whitefish MT 7,Thurmond, WV 120, Denver .2.

 

[GariYaMoshi]

The one thing I notice is that the states with more than just a single square also have more than 2 trains per day providing service.

Kind of hard for the states with just a single route, such as the Empire Builder, to get large passenger volumes when the space may already be occupied by other riders just traveling through.

As I used to teach my statistics students, there are 3 types of lies- lies, dam lies, and statistics. Learned that from my college statistics professor.

 

[jpakala]

What I wondered is why some states (e.g. VA, RI) are in boldface even though they have fewer blocks than states with more (e.g., CA, PA).

 

[Matthew H Fish]

What I wondered is why some states (e.g. VA, RI) are in boldface even though they have fewer blocks than states with more (e.g., CA, PA).

Oh, good catch. That is part of the formatting that I need to clear up.
Basically, it is just a matter of, for example, Pennsylvania having more letters than Virginia, so the font is smaller. "Virginia" isn't actually in bold, but its a bigger font size, so that makes it look bold.
When I started making this, I wasn't sure exactly how big the states were going to be, etc. That is one of the things I want to optimize.

 

[b1xn00d]

Very interesting. Large populations (Texas) and highly urbanized populations (Colorado, Utah) do not appear to be a substitute for frequent service (Delaware, Virginia).

 

[Matthew H Fish]

Very interesting. Large populations (Texas) and highly urbanized populations (Colorado, Utah) do not appear to be a substitute for frequent service (Delaware, Virginia).

That is especially apparent when comparing states near Chicago that have corridor service with more than one train a day (Michigan, Wisconsin, Missouri) with those that do not (Minnesota, Iowa, Indiana). I imagine that having the flexibility to travel to and from another city is a bigger driver of train travel than just the population density of a corridor.