And just when you thought I was done prattling on about commuter rail service patterns...
I've made a map.
First, just to get some stuff out of the way -- I am not a graphic designer and this map was adapted from an official MBTA commuter rail map using a program equivalent to Microsoft Paint. If you zoom in at all, you'll notice that it is ugly and literally a designer's worst nightmare. Originally, this was just going to be a rough proof-of-concept to plan out a nicer version. But then I got really deep into it and figured I might as well just stick with this and share it once it was complete -- hopefully inspire someone else to do it better than me.
Now, onto the map!
Project Goals
This was an attempt to visualize the peak service patterns that
I identified in my post from June. In general, I wanted to stick to the Vignelli principle of "one line, one service" so as to clearly highlight the hidden network of short-turns and expresses.
Design Philosophy
The first challenge was to come up with a design language which would capture the variability I described earlier. I eventually settled on a "most trains stop vs some trains stop" dichotomy for the station markers; combined with "one line, one service", that allowed me to capture a wide swath of the details.
Some judicious use of text helped clear most of the remaining items: first, by explicitly titling the map "Peak Services, one layer of complexity was stripped away. Second, the "very limited service" stations like Plimptonville and Silver Hill are explicitly listed out (in the same way the current map lists stations without parking): this saved me from having to create a third station marker. Finally, I added an asterisk next to Hyde Park, Route 128 and Canton Junction to call out the moderately frequent exceptions to the Providence express services there.
In several cases, my choice of categorizing a stop as "most trains" vs "some trains" on a given service is... pretty arbitrary, but overall seemed "good enough."
Exceptions to the Design Philosophy
The exceptions to the "one line, one service" principle are the Newburyport/Rockport branches, the Old Colony, and the Franklin Line. In the case of Rockburyport, both of the Long Distance services behave identically south of Salem -- this in contrast to, say, the Franklin and Stoughton Lines, where the Franklin trains mainly stop at Readville while the Stoughton trains mainly stop at Hyde Park. A Rockburyport equivalent would be Newburyport trains stopping at Lynn while Rockport trains stop at Chelsea. But they don't, so they get consolidated in order to simplify the map.
(A more rigorous application of this principle would have the Wildcat branch shown as a proper branch off of a Lowell All Stops service, instead of as a separate service, but, alas.)
The Old Colony is shown as one service because it still essentially is. The individual frequencies on each of the Old Colony Lines is markedly lower than on any other line, and you don't really get a "stacking" effect at Braintree, Quincy Center or JFK/UMass the way you do on the NEC, due to the irregularities of the schedules. (Plus, South Station was already visually overwhelming enough, with 10 terminating services.)
The Franklin/Foxboro/Fairmount Line is, unsurprisingly, the diciest part of this map. I'll go through this one by one.
The Problems With The Franklin/Foxboro/Fairmount Line
First, the Foxboro Line is visualized as out-and-out reverse-branching. I toyed with depicting Foxboro-via-Back Bay and Foxboro-via-Fairmount as separate services, but via-Fairmount service actually only has one peak trip in each direction, and its AM trip behaves identically to the via-Back Bay trips -- the PM trip is the only service to Foxboro that skips stops. That did not seem like worth cluttering the map over.
Second, the Forge Park Line... this one is a mess. According to this map, Forge Park trains sometimes skip Windsor Gardens, Norwood Depot, Islington, Endicott, and Readville. The reality is that this is (was) largely true for the PM peak -- if you wanted to go to one of those stations, you needed to take a Foxboro train, with a couple of exceptions. But the AM peak is a lot messier. During mornings, most trains originating from Forge Park do indeed stop at those stations. So it's not a hard-and-fast rule here.
What is true is that Dedham Corp Center and Norwood Central sit a head above the rest -- these are indeed the stations which every train stops at, no matter what. And that comes through, which was the key thing I wanted to convey.
In general, as I described in June, the Franklin/Foxboro Line(s) have the messiest division of labor of any corridor in the system. There's enough differentiation to make it worth considering them separate services (unlike the Old Colony), but not quite the same consistency as seen on the Reading/Haverhill or Fitchburg Lines.
Now, moving onto things that this map does well...
Successes
First, it immediately jumps out at you how the Peak Services system is really two systems superimposed -- one that is short-turn, to Beverly, Reading, Littleton, Framingham, and Stoughton, and one that is long-distance, to Rockport, Newburyport, Lowell, Wachusett, Worcester, Franklin, Rhode Island, and Plymouth County.
Second, it becomes very easy to visualize those "head above the others" stations that basically every train -- short-distance or long -- stops at, whether due to high ridership, track layout, or both. These include Salem (but surprisingly not Lynn), Reading, Anderson/Woburn, S. Acton, Waltham (sorta), West Natick, Dedham Corp Center and Norwood Central.
Third, I think this map also illustrates a number of the systemic issues with the Northeast Corridor.
First, it becomes quite clear visually that by the time Needham joins the party at Forest Hills, things are already very crowded. It's very obviously the odd one out.
Second, the aforementioned reverse branching on the Franklin/Foxboro -- while visually interesting -- is also the most visually complex feature of the map, which reflects its operational complexity.
Third, the skip-stop-esque mess that is Hyde Park, Readville and Route 128 becomes clear. There's almost nowhere else in the system that is this complex, and the visual reinforces that.
Fourth, it becomes easier to see some of the knock-on effects South Coast Rail via Stoughton would have. Stoughton serves -- like Foxboro, Framingham, Littleton, Reading and Beverly -- as a short-turn point for trains serving inner stations like Route 128, Hyde Park and Canton Junction, so that Long Distance trains from Providence and Wickford have the option to blow past them. An extension to the South Coast upsets that whole paradigm by putting Stoughton trains into the Long-Distance category, meaning they'd no longer be suitable for those inner stations. Which means you'd need to find a way to replace service to those stations -- meaning South Coast Rail isn't just a matter of extending Stoughton trains, but also means adding more trains to the Stoughton route. We can see that the NEC trunk currently supports 5 "services", which basically means you need to do some swapping. For example, swap the Needham trains into Stoughton/Taunton Short-Turns, and swap the Franklin trains into NB/FR Long-Distance.
The Unusual Case of the Providence Line
Finally -- and I think this is less obvious -- but I do think this map subtly highlights the one-of-a-kind nature of the Providence Line. It is unique in that it runs very long distance -- akin to Worcester, Fitchburg, Newburyport -- but lacks a short-turn at its midway point:
- Worcester (~43 miles) short-turns at Framingham (~20 miles, 46%)
- Wachusett (~52 miles) short-turns at Littleton (~29 miles, 55%)
- Newburyport (~36 miles) and Rockport (~35 miles) short-turn at Beverly (18 miles, 50% and 51%)
It is true that the Stoughton serves a similar purpose, but it branches off much earlier:
- Providence (~43 miles) and Wickford Junction (~62 miles) "short-turn" at Canton Junction (~15 miles, 34% and 24%)
- An equivalent short-turn would be at Mansfield (55% of the way to Providence)
The reason for this is that there's a bit of a density cavity between Canton and Mansfield, and when you look at a satellite image, you realize what's happening: the BosWash density spine heads south to Providence not by way of Canton, but via Norwood. That's where we find the stop-spacing that we see in Wellesley and the North Shore.
The result is that the Franklin/Foxboro Line essentially serves as a kind of "short-turn" service for the Providence Line, especially with the recent iteration of Foxboro trains focusing on the local stops in Norwood and Dedham. In a sense, the Foxboro Line is to the Providence Line as the Framingham Line is to the Worcester.
Conclusions
I don't think this map would be that useful for everyday riders. (Hence my hopefully impossible-to-miss disclaimers at the top.) In general, journeys on the commuter rail are between the Downtown Terminals and Someplace Else. Wayfinding just isn't as critical as on a rapid transit system, and the topology of the network just isn't that complicated. (That would change in an NSRL world, of course.)
But I think it's worth it for railfans, planners, and advocates to think of the system in these terms. The standard map belies the complexity of the network and makes it easier to avoid contending with the gnarliness of running multiple service patterns along the same route.
It also raises important questions. What are the benefits to running a system like this? What are the drawbacks? What are things we can do to simplify this system and/or to enhance it?
Thank you for reading!
