Figure 1From: An unsupervised classification method for inferring original case locations from low-resolution disease mapsPrototypical patient map for Boston, Massachusetts. The image displays 550 addresses selected by stratified random sampling design. The original JPEG image used in the analysis had a resolution of 50 dots per inch (550 × 400 pixels), a file size of 129 kb and a scale of 1:100,000. This would be a typical output for web display and usually lower resolution than would be shown in a slide presentation or in a peer-reviewed publication.Back to article page