DecisionTree - Glossary - Web-based geographic decision support and site prioritization

Adobe Flex

Adobe Flex is an open-source collection of tools that help developers make consistent, rich Internet applications, independent of an user's choice of browser. Applications made within the Flex framework-- such as the interactive map in DecisionTree-- run in Flash Player, a common browser plug-in. Flex enhances the user experience by enabling users to control graphical display elements, such as adjusting the the transparency of map layers on the fly.

API

An application programming interface (API) - is a source code interface that an operating system or library provides to support requests for services to be made of it by computer programs. In other words, an API provides a method for an application program to make requests for specific data or functionality from a remote operating system or other application.

Decision Factor

A decision factor is a particular criterion used in the DecisionTree analytical process. Each decision factor represents a geographic or demographic variable, such as proximity to public transportation or percentage of local population with a college degree. Each decision factor is represented by a single raster layer with a common extent and cell size. For analytical purposes, these layers can be combined and analyzed in a GIS using weighed overlay to provide a concrete, robust analysis.

Distributed Computing

Distributed computing enables DecisionTree to optimize the speed with which results are delivered to decision makers. As you might imagine, performing a series of map algebra calculations is a processor-intensive operation, given that the data sets in question are frequently comprised of thousands or even millions of cells. Additionally, the Web-based nature of DecisionTree demands that results be delivered to end users in a timely fashion. To this end, Azavea has exploited the nature of weighted overlay operations to develop a distributed processing methodology that optimizes resources by spreading the computing workload among multiple threads and machines.

GRID

GRID is a proprietary raster data storge format developed by ESRI. Like most raster data storage formats, Grids represent continuous, field-based attributes (such as elevation, temperature or land cover) by exhaustively tessellating space into an array of square cells. A Grid is stored in a separate directory with associated tables and files containing information about that Grid's geography, attributes and processing history. For use in DecisionTree, Grids are converted to the Azavea Raster Grid (ARG) format, which has been optimized for fast processing and storage speed.

Geoprocessing

Geoprocessing can be thought of as an operation or series of operations performed on data in a GIS.

A geoprocessing operation typically (1) ingests a spatial data set, (2) performs an analysis or data management task, and (3) outputs a new spatial dataset. These geoprocessing operations can be automated and shared through the creation of tools in scripting languages, and can be combined and sequenced to create geoprocessing models.

Heat Map

A heat map is a graphical representation of spatial data, where values are represented as colors along a spectrum, either changing hue or increasing intensity to reflect the magnitude of the variable in question. For example, areas that best meet a user's defined priorities may be colored red, while moderately suitable areas are shaded orange or yellow, progressing through green and finally to blue for the least appropriate sites. Heat maps present complex quantitative and spatial information in a form that even non-specialist decision makers can grasp. The integrated analysis produced by the DecisionTree weighted overlay process is represented as a heat map.

Map Algebra

Map algebra describes various functions and processes that can be used to manipulate raster data; it lies at the heart of DecisionTree's functionality. Map algerba enables users to combine or transform information from multiple maps by performing different kinds of mathematical functions that compute values for raster grid cells. For instance, one might calculate total spring rainfall for an area by adding several maps of monthly totals: Rain_total = Rain_March Rain_April Rain_May. For a given grid cell, the value encoded in each of the three rasters is being summed and encoded in the same location in the output grid.

DecisionTree makes power of map algebra accesible to decision makers by providing a framework in which disparate kinds of data can be integrated analytically.

Raster vs. Vector

Raster-based spatial data models define space as an array of equally sized cells (or pixels). Each cell contains an attribute value and location coordinates. Unlike a vector structure, which stores coordinates explicitly, raster coordinates are contained in the ordering of the matrix. Groups of cells that share the same value represent the same type of geographic feature. Vector-based models display and define features on the basis of two-dimensional Cartesian coordinate pairs (i.e., x and y, latitude and longitude) as points, lines, and polygons. Each point feature is represented as a single coordinate pair, while line and polygon features are represented as ordered lists of vertices. Multiple attributes can be associated with each vector feature, as opposed to a raster data model, which associates attributes with grid cells.

Scenarios

Scenarios are groups of decision factors that have been bundled together because of their relevance to particular project types. For instance, an economic development decision scenario might include zoning classes, tax abatement areas, and proximity to interstate ramps, whereas a retirement center siting scenario might include proximity to hospitals, parks and major roadways.

Spatial Modeling

Spatial Modeling can represent our world in terms of processes. In this sense, a GIS model is a sequence of processes that generate a measurement, create a map, transform existing datasets into new ones or run repeatedly to create a simulation. The objectives of a process model can vary broadly. Very commonly, a model is simply a way to automate a sequence of actions that we would otherwise have to perform manually. In other cases, the model may be generating a measurement or other output for a particular set of inputs. DecisionTree enables decision makers to generate customized spatial models by selecting and weighting decision factors to reflect their priorities.

Weighted Overlay

Weighted overlay is the technique employed in DecisionTree to combine diverse decision factors into a unified analysis. In a weighted overlay, each raster represents a particular decision factor. To combine these factors for analysis the rasters must be reclassified based on a common measurement scale, such as rankings from 1 to 9. Each raster is then weighted by the user based on the importance of that decision factor relative to others. These input rasters and their weights are then combined via map algebra operations to produce a single output raster (heat map) that represents the integrated analysis.