Introduction to Data Analysis Using Geographic Information Systems

Analytical and Operational Functions

Both traditional DBMS software and GIS support database analysis, but a GIS also supports map analysis. It is useful to think of GIS map analysis in a layered-model context. The layered GIS model is analogous to transparent maps that can be accurately stacked upon one another. Typically, each layer contains only one mapped theme. Traditionally, map analysis of multiple overlays had to be performed manually.

When attempting to answer a geographic question, the user determines which phenomena are to be examined and how the analysis will proceed. A GIS provides a set of "tools" or computer programs that allow the user to perform a specific set of operations on map and attribute data. These tools, which are in the form of operating commands, permit spatial inquiry, manipulation, and analysis. Examples of some of these operations, which give the user the power to analyze the map layers, are the focus of this section.

Functional Tools for Map Analysis

Tools that manipulate attribute data may employ logical expressions similar to those described in the database analysis discussion presented above. In terms of map data analysis, and with the exception of those tools used specifically for overlaying (or merging) data from multiple map layers, operations performed on map data can be performed on either a single layer or upon multiple layers.

The functional tools available for map analysis can be grouped into six categories: (1) projection and spatial transformation utilities, (2) spatial retrieval, classification, and measurement functions, (3) logical and visual overlaying capabilities, (4) proximity and network functions, (5) map algebra utilities, and (6) output generation.

While the specific names for the various analytical tools may vary from one GIS to another, the operations they perform are similar. The tools may be used in various combinations and sequences to accomplish the desired task. The selection and sequencing of tools should be determined according to the specific need at hand. Some of the spatial analysis operations that can be performed using these tools are illustrated in Figures 3 - 7.

Projection and Spatial Transformation Utilities

Coordinate change Projection Edge Matching Figure 3. Projection and spatial transformation utilities allow a GISto reconcile irregularities between map layers. Map data can be mathematically converted from one coordinate and projection system to another.

Various pre-processing operations may need to be performed to remove errors that may occur as observations are made, maps are compiled, and/or as layers are encoded into the database. A common source of error is in the registration of multiple layers. Registration involves the systematic adjustment of a map layer so that it can be accurately laid over another layer of the same area. Data development may not consistently use the same units of measure during data collection. Because of differences in data sources, errors during data input, and because some layers may be in one coordinate system [such as Latitude/Longitude or Universal Transverse Mercator (UTM) units] while some may be in another, registration is important. A GIS is able to reconcile these irregularities by mathematically converting the map data from one coordinate and projection system to another (Figure 3).

Numerous map projections have been developed to represent the earth's three-dimensional surface on a two-dimensional map. Usually, a GIS user has no control over the projection of existing maps. Some maps may be in a cylindrical Mercator projection, others in a conic Albers, and still others in a planar projection. A typical GIS will have a set of programs which contain the algorithms needed to convert data from one map projection to another (Figure 3).

Another problem commonly arises where the area of interest is on adjoining maps. Here the location of some geographical objects may need to be modified to create a "seamless" layer between maps. This process requires that objects spanning two adjoining maps be "edge matched" to reconcile shape and location discrepancies (Figure 3). Often the margin of error is distributed between the adjoining polygons or lines. For example, a road layer may span several maps. Assume that one road is disjointed as it crosses into an adjoining sheet. The operator can reconcile the difference and make the road continuous by editing both parts of the study area simultaneously and removing the error (bringing the separate road segments together at a mutual point).

Spatial Retrieval, Classification, and Measurement Functions

Figure 4. Spatial retrieval, delineation and classification, and measurement are separate functions, but are commonly used together. For example, a GIS user interested in legislative reapportionment may want to retrieve a district map, classify zones according to party affiliation, and determine the size of the area under study.

In the previous discussion of database analysis, a GIS was asked to locate lakes according to user-specified criteria and the retrieved data was presented in a tabular format. In contrast, data may be retrieved according to its location on a map layer or its spatial relationship to other mapped data. This process is known as spatial retrieval (Figure 4). Spatial retrieval produces a spatial representation (map). For example, polygons representing lakes between 0 and 10 acres could be coded in one color, those between 11 and 20 in another color, and so forth, and the resulting color map could show the distribution of the various-sized lakes. The polygons could be presented in a spatial pattern corresponding to the actual location and configuration of the lakes in the real world. This spatial representation would likely be more meaningful to the user than the same information presented in a tabular format.

When map data are graphically displayed on a computer monitor, it is possible to select a specified area and examine it visually or analytically in more detail. This is known as "windowing" or "clipping" (Figure 4). This may be done by physically indicating the desired location or by entering coordinate values that delineate the area of interest. The elements (polygons, lines, or points) found within the windowed area can be extracted and set aside on a new layer, allowing faster analysis of the smaller, selected data set. During windowing, the scale of the display can be changed to allow a more detailed display of the area of interest (provided that detailed data for the area of interest is present in the database). Conversely, the user may wish to generalize data and window "out" to a less-detailed display.

Classification of spatial phenomena requires that the many types of data commonly found in maps, aerial photos, or satellite imagery be interpreted and coded so that it can be stored and used in the GIS (Figure 4). The GIS allows the captured data to be manipulated and combined until the appropriate aggregation occurs. For example, zones of poverty may be defined by classifying and combining data on unemployment, government assistance payments, and income statistics. By combining these layers, patterns may be delineated. Those patterns can also be subjected to further classification schemes, if necessary (e.g., setting priorities for the establishment of social service offices).

Determining the length of a road or river, the area of a village, or the density of a population are all measurement problems (Figure 4). A GIS allows both simple and complex measurement functions to be performed. By automating measurement tasks that are typically performed manually, the GIS frees the user to perform other analyses or tasks that require more skill or involvement.

As mentioned earlier, a GIS allows the conversion of spatial data in one unit of measurement to that of another. Lengths captured in meters may be converted to feet, acres to square miles or hectares, etc. Vertical and slope distances are commonly produced when exploring three-dimensional relationships in digital elevation models. Bearings, or azimuths, and other survey data may be transformed to yield maps specifying locations, distances, and precise angles.

The GIS user may need to know how many times a spatial phenomenon occurs. For example, someone studying contaminated groundwater well sites may want to know the location of wells that become contaminated and may also want to know the number of contaminated sites. In this case, the user will want map data that can be used to indicate both spatial patterns and frequency of occurrence.

In certain spatial analyses, the user may want to find the best division of space based on spatial and non-spatial attributes. Attribute data within the study area are ranked, correlated, and then used to assist in the spatial ordering and delineation of areas. Legislative reapportionment is an example where spatial attributes (location and population) may be weighted with a non-spatial attribute (political bias) to calculate areal divisions.

Logical and Visual Overlaying Capabilities

Overlay Merging Figure 5. The overlay function allows the user to "stack" map layers on one another, showing spatial relationships between the layers. Merging generalizes classes within map layers by combining attributes to reveal new map features.

GIS is probably best known for its ability to build map layers and evaluate the relationships between them. The relationships between map layers can be assessed from both mathematical (logical) and graphic (visual) perspectives.

Logical overlays superimpose layers using logical or spatial functions and store the results in the GIS database as new layers of data. This mathematical approach examines the quantitative association between the phenomena of interest. This relationship is determined by combining various data layers to create a composite data set (Figure 5). Because the layers have already been registered, they can be accurately placed over one another. In a typical GIS, the analysis might require that data from existing layers be combined to create new data or map layers. For example, streams (lines) and deer sightings (points) can be combined with forest stands (polygons) to create a new database that reveals the spatial relationships between all three in the form of habitat quality.

The merging of data allows the user to take a complex data layer and to dissolve lines between shared attributes (Figure 5). The result is a more general data layer. Merging is the reverse of the process of combining various data layers or classes of attributes, as described above. As an example, if an ownership layer contains data for state government, local government, nonindustrial private owners, and private industry, a merge could be performed to reveal the spatial relationship between public and private ownerships.

Visual overlays, on the other hand, allow the user to graphically view spatial relationships between the various layers instead of seeking specific mathematical relationships. The overlays are graphically presented on a map or computer monitor. Visually displaying these overlays does not create a new layerin the database; it simply provides visual cues of the relationships between the layers.

Proximity and Network Functions

Buffering Networks Figure 6. The buffer function examines an area which surrounds an object of interest. This function is used to create zones and to determine routes within zones. For example, a zone can be created based on specified distances from map features (e.g., the area within five miles of a road). Network functions examine the movement of objects along an interconnected pathway (e.g., traffic flow along a map of highway segments).

These types of analyses consider predefined areas around a geographical object or the connectivity of phenomena (Figure 6). For example, assume that the user had two layers, one of roads and another of timber stands with average stand diameter as an attribute. Furthermore, assume that the user wanted to know how much merchantable timber (trees of a specified diameter) was located within 500 feet of any road. By setting the proper parameters that consideredmerchantability and road location, a buffer zone would be created measuring 500 feet on either side of each road. All appropriate map and attribute data pertaining to the timber resources within the buffer would be generated as a separate layer available for observation and analysis.

Another type of proximity analysis involves the network function of a GIS (Figure 6). Networks are commonly established to evaluate options for the purpose of route optimization and resource allocation. Specifically, this means locating the best route between two points or the selection of service zones in a network (e.g., pizza delivery areas, fire service zones, mail routes). A common network function is the routing of emergency vehicles on road systems. In this case, the GIS analyzes distance factors, road speed, and other transportation variables (e.g., flow of traffic, traffic control measures) to generate alternative routes between two points.

Map Algebra Utilities

Figure 7. Map algebra utilities allow the user to specify mathematical relationships between map layers. Entire maps can be added, subtracted, multiplied, and divided according to user-specified rules. For example, a new map can be generated by determining the difference in elevation between a topographic map and the corresponding map of the water table.

Another very useful, quantitative capability of GIS is the application of algebraic expressions to map layers (Figure 7). Referred to as map algebra, this process enables users to specify mathematical relationships between map layers. Thus, entire maps can be added, subtracted, multiplied, and divided, according to user-specified rules. As an example, a new map can be generated by determining the difference in elevation between the topographic map and the map of the water table. The resulting layer can then be stored for subsequent analyses.

Output Generation

The final set of tools we consider here provide the ability to create output such as maps, geographical summaries or reports, and geographical base files (files containing both the digital map and attribute data). Output can be either hard copy, digital files, or displayed on a computer monitor. The sophistication of the GIS software and the output capabilities of the hardware/software system dictate the quality and variety of options available to the operator. In most cases, the GIS allows maps, summaries, and base files to be written in a number of different digital export formats so that they may be used by another GIS. Most systems have the ability to translate or directly import and export these files.

Generating graphic output, commonly in the form of maps, requires that a GIS have a wide variety of symbols and format options. Most offer numerous line, polygon, and point symbols to represent geographical phenomena as well as text options for labeling and annotating output. Since these utilities allow maps to be produced at various page sizes and map scales, output can be custom-designed to a format that is most appropriate for the situation.

Geographical and tabular summaries are common types of GIS output. Such summaries differ from those created by a traditional database query because they are based on map analysis. For example, upon completion of a windowing of clipping function, summary statistics presented in tabular form for the clipped area are often necessary to complement the analysis. A GIS will usually offer this type of report-generating capability and it is especially useful during complex spatial analysis or database documentation.

Use of reference maps is an important part of the output (mapping) process. It is common for a GIS to contain a library of basic, often-used map and attribute data of the study area for creating these simple reference maps. Such base files could include county, section, parcel, or zip code boundaries, major transportation routes, or hydrological features. When included in output maps, these cartographic features provide a useful frame of reference for the map user.

Produced by Communication and Educational Technology Services, University of Minnesota Extension.

In accordance with the Americans with Disabilities Act, this material is available in alternative formats upon request. Please contact your University of Minnesota Extension office or the Extension Store at (800) 876-8636.

University of Minnesota Extension is committed to the policy that all persons shall have equal access to its programs, facilities, and employment without regard to race, color, creed, religion, national origin, sex, age, marital status, disability, public assistance status, veteran status, or sexual orientation.