Tuesday, February 23, 2010

Overgeneralizing About Bad Transit

Source: my analysis of local bus service frequencies listed on a LACMTA transit map

I have a hypothesis: some people have had a bad experience on transit and come to the conclusion that all transit experiences are bad experiences. This determines their opinion about the viability of transit as a mode of transportation in general.

Of course this is a fallacy of reasoning, but, considering that America's landscape is predominantly suburban and transit has trouble without supporting density, this hypothesis could be a cause for concern.

To take an example, the main bus that goes by where I grew up comes by only once every 45 minutes. Say, in a rookie attempt to use it, I just went out to the nearest bus stop without looking up the timetable. Imagine further that I just missed the previous bus and had to wait the entire 45 minutes for the next one. That sucks. I could have easily overgeneralized from that bad experience.

A look at Metro's timetables for regular buses shows that service frequency varies widely. For weekdays during the day I evaluated service frequency on "local buses" and came up with categories based on how long one has to wait between buses: "excellent/A" (10 minutes or fewer), "good/B" (11 to 20), "fair/C" (21-30), "poor/D" (31-45), "abysmal/F" (46-60), and "soul crushing/F-" (over an hour).

You can see in the graph above how Metro's local service stacks up (remember, this only includes the regular buses with frequent stops, the bread and butter of LA's transit system). Metro's GPA isn't too hot, and if the bad transit overgeneralization hypothesis is correct, we should be very worried about that.


  1. Heh... don't do that analysis on RTA. Our two best routes are "fair" (20 and 30), we've got one that comes in at "poor" (40), and then a few "abysmals" and a whole lot of "soul-crushing". By your reckoning, the souls of every RTA rider who doesn't live in select areas of Riverside, Moreno Valley and a tiny slice of Corona is having their souls crushed on a daily basis.

    (Interestingly, if you combine the two routes that serve my area, we get a B. And a decent C until half past midnight, four nights a week.)

  2. A pie chart could work too. I've had a teacher who hates them though so I'm out of the habit of using them :)

  3. I don't know why your teacher hates them; they convey the fact that these are all percentages of the same total, much better than the bar graph.

  4. His argument against pie charts is that if you want to compare two periods of time (or more), putting two pie charts next to each other is harder to read than putting two bars next to each other in different colors.

  5. Oh sure. But if you're just doing a single point in time, pie charts are better than bar graphs. For some jobs you want a hammer, for others a screwdriver.