Ghost Map

Steven Johnson's new book: Ghost Map: The Story of London's Most Terrifying Epidemic ..., re-tells the story of how the water-borne model of cholera epidemics replaced the miasma (bad smell) model in the 19th century and effectively led to the elimination of cholera in the developed world. Like many non-epidemiologists, I first heard of this story via Edward Tufte's 1983 book: The Visual Display of Quantitative Information. I was then more interested in the visual aspects of how the map of the epidemic made the case.

This book tells a more complete story, following the reasoning and collaboration of John Snow and Henry Whitehead which ultimately led to the change in understanding of cholera epidemics, which resulted in the re-engineering of city sewage and water systems.

It turns out that Tufte made a number of errors in his description of the epidemic in his first book, and largely corrected them in his followup book: Visual Explanations. Even there he misses some of the subtlety of the maps, for example their use of Voronoi diagrams to model the proximity of people to the source of infection. Tufte does an excellent job in understanding the effectiveness of visual representation, but less well with integrating computational methods that can aid human understanding.

Also notable is Johnson's hypothesis that this discovery was one of the first triumphs of amateurs. Snow was a physician, Whitehead was a local clergyman. It was their use of local information, painstakingly gathered, that ultimately made the case against a medical community entrenched in an incorrect model. Taking this further, the existence of the Web now makes it much easier to collaborate, gather data and drive to alternative solutions. Or does it confuse the matter through all the clutter?

The mapping itself today can be readily and cheaply done via approaches like Google Earth. The inclusion of such 'local information' as the proximity of pizza parlors can also be used to model complex visual epidemic data. Will this be the value of desktop visualization?

In the last chapter of the book Johnson examines the architecture of cities and the implications of how their scale effects such challenges as epidemics and nuclear terrorism. He is more speculative here, but the material is thought-provoking.

Excellent book for those interested in the history of this episode, the practical use of visualization and how difficult it is to change existing models.