Thanks to everyone (and I do mean everyone!) who has sent the recent “The World’s Most Complex Subway Maps as Determined by Scientists!” article to me – from various sources, including this take from CityLab. 

However, when I read the full academic paper that all these articles are based on, I think that everyone’s got the wrong end of the stick. The study is not of map complexity at all, but of network complexity. 

The methodology outlined at the end of the paper makes it very clear that a theoretical topological network has been assembled for each city based on information from Wikipedia and data feeds from the relevant transit agencies, with travel and transfer times accounted for within each model. These models are then tested mathematically – entirely with equations – to determine the complexity level of each system. 

At no point is an real official printed map used, nor are there any usability tests performed by real humans. It seems to me that the actual design of a map – which can make a simple network incomprehensible or a complex one easy to navigate – is not considered at all in this study. So while the study is interesting, and reveals a lot about the maximum amount of information that a human can reasonably hope to remember (the start point, two interchange points and the end point, basically), it really doesn’t say anything about how map design can help or hinder that process. 

In other words: According to the study, New York has the most complex transit network in the world, but not necessarily the most complex map.

Source: ScienceAdvances Journal

Sidenote: If you’d like to read a proper usability test paper for transit maps, check out this one that Max Roberts did for the Paris Métro, comparing “standard” octolinear maps to his curvilinear version. [PDF]