and county boundaries and lakes), a line layer of rivers, and a point layer of cities that have populations of 250,000 or more. The state and county boundaries are coterminous — that is, they share boundaries and do not overlap each other. Color fill is used within the county polygons to show the mortality of lung cancer for white males. This map has some striking geographic patterns that are discussed later in this chapter.
Associated with individual point, line, or polygon features are data records that provide identifying and descriptive data attributes. For example, in figure 1.1, the labels for the names of states and cities come from tables of attribute records associated with each map layer. You will revisit this map in tutorials 1-3 and 1-4 in this chapter where you will use ArcGIS to explore map layers and spatial patterns of cancer mortality.
Figure 1.1 Lung cancer mortality per 100,000 white males, 2000 – 2004.
Sources: (a) Esri Data & Maps; (b) Cancer Mortality Maps website, National Cancer Institute.
Raster maps are stored in standard digital image formats, such as tagged image file format (TIFF) and Joint Photographic Experts Group (JPEG) files. An image file is a rectangular array, or raster, of very small, square pixels. Each pixel, or cell, has a single value and solid color and corresponds to a small, square area on the ground, from 6 in. to 3 ft on a side for high-resolution images. Accompanying the image files are world files that provide georeferencing data, including the upper-left pixel’s location coordinates and the width of each pixel in ground units. Using world file information, GIS software can assemble individual raster datasets into larger areas and overlay them with aligned vector datasets.
Viewed on a computer screen or on a paper map, a raster map can provide a detailed backdrop of physical features. In figure 1.2, an aerial photograph overlaid with vector map layers shows locations where serious injuries of child pedestrians occurred in relation to public parks that have playgrounds. The two boundaries surrounding the parks are 600 ft and 1,200 ft buffers used to study injury rates near parks. You will explore the GIS data behind this map in depth in chapters 4 and 7, where you will download similar orthoimagery and create and use buffers similar to those in figure 1.2.
Figure 1.2 Locations of serious injuries to child pedestrians in eastern Pittsburgh, Pennsylvania.
Sources: (a) Children’s Hospital of Pittsburgh; (b) US Geological Survey; (c) City of Pittsburgh, Department of City Planning.
Map layers have geographic coordinates, projections, and scale. Geographic coordinates for the nearly spherical world are measured in polar coordinates, and the angles of rotation are measured in degrees, minutes, and seconds, or decimal degrees. The (0,0) origin of the coordinate system is generally taken as the intersection of the equator and the prime meridian (great circle), which passes through the poles and an observatory in Greenwich, England. Latitude is measured north and south for up to 90 degrees in each direction. Longitude is measured to the east and west of the origin for up to 180 degrees in each direction.
The world is not quite a sphere because the poles are slightly flattened and the equator is slightly bulged out. The world’s surface is better modeled by a spheroid, which has elliptical cross sections and two radii, instead of the one radius of a sphere. The mathematical representation of the world as a spheroid is called a datum; for example, a datum commonly used for North America is North American Datum of 1983 (NAD 1983). If you use the same projection but two different datums for your maps, each corresponding map will have small but noticeable differences in location.
A point, line, or polygon feature on the surface of the world is on a three-dimensional spheroid, whereas features on a paper map or computer screen are on a flat surface. The mathematical transformation of a world feature into a flat map is called a projection. There are many projections, some of which you will use in chapter 4. Each projection has its own rectangular coordinate system and a (0,0) origin conveniently located so that coordinates are positive and have distance units, usually in feet or meters.
All projections necessarily cause distortion of direction, shape, area, or length in some combination. So-called conformal projections preserve shape at the expense of distorting area. Some examples are the Mercator cylindrical and Lambert conic projections. Equal-area projections are the opposite of conformal projections: They preserve area while distorting shape. An example is the Albers equal-area projection (Clarke 2003, 42 – 44).
Map scale is often stated as a unitless, representative fraction; for example, 1:24,000 is a map scale where 1 in. on the map represents 24,000 in. on the ground, and any distance units can be substituted for inches. Small-scale maps have a vantage point far above the earth and large-scale maps are zoomed in on relatively small areas. Distortions are considerable for small-scale maps but negligible for large-scale maps relative to policy, planning, and research applications.
GIS maps are composites of overlying map layers. For large-scale maps such as figure 1.2, the bottom layer can be a raster map that has one or more vector layers on top, placed in order so that smaller or more important features are on top and not covered up by larger contextual features. Small-scale maps, such as figure 1.1, often consist solely of vector-map layers. Each vector layer consists of a homogeneous type of feature — points, lines, or polygons.
Digital map infrastructure
GIS is perhaps the only information technology that requires a major digital infrastructure. The map layers of the infrastructure are referred to as basemaps — namely, a collection of standards, codes, and data designed, built, and maintained by government. Vendors provide valuable enhancements to the digital map infrastructure, but for the most part, it is a public good financed by tax dollars. Without this infrastructure, GIS would not be a viable technology.
The National Spatial Data Infrastructure (NSDI), developed by the Federal Geographic Data Committee (FGDC at http://www.fgdc.gov), incorporates policies, standards, and procedures that allow organizations to produce and share geographic data. The Geospatial One-Stop website (http://geo.data.gov/geoportal), which is part of the NSDI, provides access to spatial data. Websites change often. The websites noted in this book could be named differently or redirected to different sites, but the information will still be available if you search for it. GIS provides a way to use this wealth of publicly available data in your own studies.
Perhaps the most useful spatial data for health applications comes from the US Census Bureau in the form of TIGER/Line maps. These maps are available by state and county for many classes of layers. These classes, and examples of each, include the following:
• Political: states, counties, county subdivisions (towns and cities), and voting districts
• Statistical: census tracts, block groups, and blocks
• Administrative: ZIP Codes and school districts
• Physical: highways, streets, rivers, streams, lakes, and railroads
You can download TIGER/Line map layers in GIS-ready
