Tessa Holloway has picked up on a curious anomaly about the pattern of Vancouver’s street grid in a rather technical academic paper: “A typology of street patterns.

Remi Louf and Marc Barthelemy of France’s Institut de Physique Theorique developed “a quantitative method to classify cities according to their street pattern (to) define what could constitute the ‘fingerprint’ of a city.”


Block patterns


They grouped block patterns into four groups (above) – and mapped the results:

Street grid

If you check out North America, you can see all the cities are yellow dots (Group 3, the predominant global pattern) except Vancouver (and Mexico City, which the authors don’t seem to include in North America).

Although one should be wary of sampling bias here, it seems that the type of pattern characteristic of the group 3 (various shapes with larger areas) largely dominates among cities in the world. Interestingly, all North American cities (except Vancouver, Canada) are part of the group 3, as well as all European cities (except Athens, Greece). The composition of the other continents is more balanced between the different groups.

Vancouver, a Group 2 city, has “a dominant fraction of small blocks with shapes broadly distributed.” Group 3 is “similar to the group 2 in terms of the diversity of shapes but is more balanced in terms of areas, with a slight predominance of medium size blocks.”

Tessa has these observations:

What jumps out to me is that the difference between Group 2 and Group 3 seems to lie in the size of the city blocks; in particular, Group 2 has smaller average block size.

I wonder what impact this has on walkability, seeing as a more connected street grid implied by smaller block sizes tends to support high walkability scores. Does that mean Vancouver, the only Canadian or U.S. city included in the study to be in Group 2, has an advantage regarding walkability that other cities don’t?

I also wonder, though, how different Group 2 and Group 3 truly are. The example cities in the study – Group 2’s Tokyo and Group 3’s New York – appear quite different regarding block size, but Seattle and Vancouver always appeared to me, at least subjectively, as having similar street patterns. I wonder if this study has uncovered a difference by looking at the blocks differently, but it could also be that the data is influenced by other things, such as whether suburban areas are included in the city’s borders.

I also wonder how the data would be different if you break down a city into different neighbourhoods. The study authors did this for New York’s five boroughs and found striking differences, but nowhere else. Downtown’s Portland square blocks appear similar to Buenos Aires’s street grid, which got its own category as Group 1, but while Portland’s street grid morphed as the city grew outwards, Buenos Aires is much more uniform across the city. Maybe breaking up the city into different neighbourhoods will reveal a great deal more about each city than using the whole city’s boundaries.


The study authors claim they’ve found a way to study street patterns objectively, and it’ll be interesting to see if this work proves useful or if we can find patterns in the data that point to particular outcomes, or use a similar method to uncover more categories.


UPDATE: Sam Sullivan, who takes a keen interest in how cities are shaped, adds this observation:

Vancouver is lucky it didn’t get caught in the Dominion Land Survey and just escaped Col. Moody’s version of the American Public Land Survey because he focused on the Fraser Valley and New Westminster. Many of the District Lots in Vancouver were surveyed separately which could have something to do with it. Also (CPR surveyor) Lauchlan Hamilton didn’t use 66-foot Gunter’s chain but Ramsden’s 100-foot chain in the CPR areas, which led to narrower lots.

District lots in Vancouver:

District  lots


Land survey South of the Fraser: