Tools for making sense of spatial variation

In this session, we will focus in on the problem of how we can use epidemiological data to examine the role of Tobler’s first law - the universality of spatial relatedness - as well as Goodchild’s revised first law - the ubiquity of spatial variation - in epidemiology data.

Before Class

Please read the following two selections. The first is a short article introducing the concept of a disease ‘hotspot’ and the second is a more in-depth treatment of the idea of disease clustering:

  1. Read the piece “What is a Hotspot, Anyway?” by Lessler et al. (2017)

  2. And Chapter 3: Interpreting Clusters of Disease Events from (Lawson et al. 2016)

During Class

During the first half of class, we will discuss the readings and key concepts in them and I will introduce the concept of kernel smoothing and its application to epidemiologic data as a tool for visualizing and identifying disease clusters.

In the second half of class, we will complete a guided tutorial going over the basics of locally-weighted regression and, if we have time, a hands-on activity in which we will experiment with smoothing different types of simulated data using this interactive tutorial.

Additional Resources


Zoom recording of class session


Lawson, Andrew B., Sudipto Banerjee, Robert P. Haining, and Maria Dolores Ugarte, eds. 2016. Handbook of Spatial Epidemiology. 1st edition. Boca Raton: Chapman and Hall/CRC.
Lessler, Justin, Andrew S. Azman, Heather S. McKay, and Sean M. Moore. 2017. “What Is a Hotspot Anyway?” The American Journal of Tropical Medicine and Hygiene 96 (6): 1270–73.