Taking the bus

Bert Gunter writes:

This article on bus ridership is right up your alley [it’s a news article with interactive graphics and lots of social science content].

The problem is that they’re graphing the wrong statistic. Raw ridership is of course sensitive to total population. So they should have been graphing is rates per person, not raw rates. I grant you that maybe populations didn’t change that much over the time concerned — but I don’t know that! At least they should have said something about the necessity of that assumption and possible distortions it could cause in their “analysis.” Note that the ensuing discussion in the article speculated about explanations for a decline in the individual propensity to ride (financial, age,…) for which the per person frequency should be the basis, not the per city.

My reply: I agree that they should’ve divided by population. Actually, I think best would be to divide by population * days, so that what they’re plotting is average number of bus rides per person per day. That’s a number that is directly interpretable. For example, 0.1 bus rides per person per day corresponds to 10% of the people in the city riding the bus once that day, or 5% riding the bus twice. The number is still only approximate, as suburbanites and visitors ride the bus too, but it’s at least roughly interpretable.

This sort of thing comes up a lot, the value of rescaling statistics to be on the human scale. Such rescaling is not always so easy—for example, reporting suicide rates per 100,000 is pretty much uninterpretable, but it’s not clear how to put such a rare event on unit scale—but we should do this when we can.

It also looks like they screwed up on the above graph and there was some software setting that cut off the lines when they went below -13%.

My other comment is about the bus riding experience itself. The authors talk about buses vs. trains and bikes and cars, but I don’t see anything about what it feels like to actually ride the bus, except for a brief mention of bus lanes. I like riding the bus, but not when it stops every two blocks, when it has to swerve in and out of traffic after each bus stop, when it stops at just every traffic light, when I have to wait 20 minutes for the bus to show up in the first place (particularly annoying because if you’re not staring eagle-eyed at the street, the bus might come by and not stop for you), etc etc. Lots of these problems are potentially fixable, for example by putting bus stops in the middle of the street and running smaller buses more frequently rather than huge buses further spaced in time.

10 thoughts on “Taking the bus

    • A while ago, I read an article discussing missing bus trips due to drivers not showing up for work and not telling anyone that they wouldn’t be there in Boston. The numbers were such that it sounded as though there were more missed bus trips per week in Boston than per year in Tokyo*, and Tokyo has a lot more busses than Boston. Here, the waits between bus are sometimes long, but if the sign says there will be a bus at such and such a time, there will be one.

      The worst thing about buses in Tokyo is that every stop has a name and you can’t read the names of the stops unless you already know the answer**. (The other Tokyo joke is that you can’t deliver mail unless you grew up in the neighborhood you are delivering mail in. Addresses are logical and hierarchical, but the spatial arrangement of things at each level in the hierarchy is random. (For example, buildings in a block are numbered in the order they were built.))

      Also somewhat related to discussions below, despite going on semi-lockdown, Tokyo retained all train and bus schedules, so that people who did have to go in to work would at least see the advantages of reduced ridership.

      *: My initial guess was per day in Boston vs. per decade in Tokyo. But not having the numbers, I’ll be more conservative.
      **: The Japanese like Chinese characters even more than the Chinese do, and have great fun assigning new and arbitrary meanings and pronunciations to characters just for the fun of it. And using them for place names.

  1. The NYT article

    https://www.nytimes.com/interactive/2020/03/13/upshot/mystery-of-missing-bus-riders.html

    is dated March 13, 2020 so presumably it does not reflect the evolving influence of the Covid-19 virus.

    With regard to Minneapolis, the article states, “Some bus trips appear to have shifted to a newly opened light rail line.” But light rail and public bus traffic in Minneapolis are an integrated unit so the loss in bus trips is merely a reflection of bookkeeping.

  2. Here is a new preprint by some of my colleagues on the impact of COVID-19 on transit ridership in Nashville and Chattanooga TN. https://arxiv.org/abs/2008.02413

    “we provide a data-driven analysis of COVID-19’s affect on public transit operations and identify temporal variation in ridership change. We then combine spatial distributions of ridership decline with local economic data to identify variation between socio-economic groups. We find that in Nashville and Chattanooga, TN, fixed-line bus ridership dropped by 66.9% and 65.1% from 2019 baselines before stabilizing at 48.4% and 42.8% declines respectively. The largest declines were during morning and evening commute time. Additionally, there was a significant difference in ridership decline between the highest-income areas and lowest-income areas (77% vs 58%) in Nashville.”

    • “…there was a significant difference in ridership decline…”

      Might want to be careful with the infamous “s” word. It’s not totally clear if it’s the “S” word (e.g., statistical significance) or the “s” ( = notable difference) word. It looks like the latter, but maybe it’s both?

  3. When checking the NYT original article I cannot spot the mentioned truncation at 13%, so it may be an artifact of the screenshot. Beyond this factual comment, I do not read much from the article, which seems to dump all sorts of “explanations” into the analysis. The never-ending expansion of cities is mentioned albeit being another cause for not taking the bus, although the time scale may be too short for this to be a relevant cause.

  4. From the article:

    > Sometime around 2013, bus ridership across much of the country began to decline.

    Why 2013? Eyeballing the graph, the across-the-board downturn began in 2016.

Leave a Reply to Mendel Cancel reply

Your email address will not be published. Required fields are marked *