Geographic Information Science and Systems

GIS as a system, science, and area of study

Longley et al. open with a crucial distinction: GIS can be understood as a system (software + hardware + data), a science (the study of fundamental spatial questions), and an area of study (academic discipline). This tripartite framework helps practitioners position their work within the broader intellectual landscape.

GIScience, as coined by Michael Goodchild in 1992, goes beyond the technology to ask: How do we best represent geographic reality in digital systems? How do spatial patterns emerge? How does uncertainty propagate through spatial analysis? These questions have no simple answers and drive ongoing research in spatial cognition, computational geometry, and spatial statistics.

The chapter introduces the concept of spatial thinking — the ability to reason about location, distance, direction, and spatial relationships. Research shows that spatial thinking is a distinct cognitive ability that can be developed through training, and GIS is one of the most powerful tools for developing this skill.

• GIS operates as system, science, and field of study simultaneously
• GIScience asks fundamental questions about spatial representation
• Spatial thinking is a trainable cognitive ability enhanced by GIS

Cross-industry GIS applications

This chapter showcases the breadth of GIS applications through real-world case studies spanning government, business, environment, and social sectors. Examples include: John Snow's cholera map (1854) demonstrating early spatial epidemiology; Walmart's use of GIS for site selection and supply chain optimization; and conservation organizations using GIS to track endangered species habitat.

The authors categorize applications by scale (local, regional, national, global), sector (public, private, nonprofit), and function (inventory, analysis, planning, management). This taxonomy helps readers identify where GIS adds the most value in their own domains.

• GIS applications span virtually every industry and scale
• Value comes from combining location with domain knowledge
• Both public and private sectors derive significant ROI from GIS investments

Conceptual models of geographic reality

How do we translate the continuous, messy, infinitely complex real world into neat digital categories? This chapter explores the fundamental conceptual models that underpin all GIS data: the field model (geography as continuous surfaces — elevation, temperature, pollution concentration) and the object model (geography as discrete, identifiable things — buildings, roads, administrative units).

The Modifiable Areal Unit Problem (MAUP) receives detailed treatment. MAUP demonstrates that statistical results change when the same data is aggregated to different spatial units. Census data summarized by zip code produces different patterns than the same data summarized by census tract. This is not a bug but a fundamental property of spatial data that analysts must acknowledge and manage.

Scale is examined not just as a map concept but as a fundamental organizing principle of geographic knowledge. Features that are visible at one scale disappear at another. Processes that operate at one scale may be driven by different mechanisms at another. Understanding multi-scale geographic thinking is essential for proper GIS analysis.

• Field and object models represent fundamentally different views of geography
• MAUP means statistical results depend on spatial unit choice
• Scale affects what is visible, measurable, and meaningful in spatial analysis

Spatial data quality, uncertainty, metadata

All geographic data contains uncertainty — in position, attributes, time, and classification. This chapter provides a rigorous framework for understanding, measuring, and communicating data quality. The authors distinguish between accuracy (closeness to truth), precision (reproducibility of measurements), and resolution (finest distinguishable detail).

Uncertainty propagates through GIS operations. When two uncertain layers are overlaid, the resulting layer inherits and often amplifies the uncertainty of both inputs. Monte Carlo simulation and sensitivity analysis are presented as methods for quantifying propagation effects. The chapter emphasizes that ignoring uncertainty does not eliminate it — it merely hides it from decision-makers.

Metadata — structured data about data — is the mechanism for communicating quality information. Standards like ISO 19115 define what metadata should contain: lineage (how the data was created), positional accuracy, temporal extent, completeness, and logical consistency. Without metadata, spatial data is unreliable regardless of its apparent quality.

• All geographic data contains uncertainty that must be acknowledged
• Uncertainty propagates and amplifies through spatial analysis operations
• Metadata is essential for evaluating data fitness for use
• Accuracy, precision, and resolution are distinct quality concepts

Architecture, databases, standards

The GIS software landscape encompasses desktop systems (ArcGIS Pro, QGIS), server platforms (GeoServer, ArcGIS Enterprise), cloud services (Google Earth Engine, ArcGIS Online), and database systems (PostGIS, Oracle Spatial). Understanding this architecture helps organizations select appropriate platforms for their needs.

Spatial databases extend relational databases with geometric data types and spatial indexing. PostGIS, the spatial extension of PostgreSQL, supports ISO SQL/MM spatial types, R-tree indexing for fast spatial queries, and over 300 spatial functions. Enterprise GIS increasingly relies on spatial databases rather than file-based data storage.

Interoperability standards from the Open Geospatial Consortium (OGC) enable different GIS systems to share data. Key standards include WMS (Web Map Service) for map images, WFS (Web Feature Service) for vector features, WCS (Web Coverage Service) for raster data, and GeoJSON for lightweight web data exchange.

• GIS architecture spans desktop, server, cloud, and database tiers
• Spatial databases provide enterprise-grade data management
• OGC standards enable interoperability between different GIS platforms

Spatial statistics, modeling, optimization

Spatial analysis goes beyond simply looking at a map — it applies mathematical and statistical methods to detect patterns, test hypotheses, and build predictive models. This chapter covers spatial autocorrelation (Moran's I, Geary's C), hot spot analysis (Getis-Ord Gi*), spatial regression, and geostatistical interpolation (Kriging).

Tobler's First Law of Geography — "everything is related to everything else, but near things are more related than distant things" — is the foundation of spatial statistics. Most spatial datasets violate the independence assumption of classical statistics because nearby observations are correlated. Spatial statistics methods account for this autocorrelation.

Optimization models use GIS to find best solutions for spatial problems: optimal facility locations (p-median, p-center models), efficient routing (traveling salesman, vehicle routing), and resource allocation (maximizing coverage with limited resources). These methods combine GIS data with operations research algorithms.

Multi-criteria decision analysis (MCDA) integrates multiple spatial factors into a single suitability map. Weighted overlay analysis assigns importance weights to different criteria (soil quality: 30%, slope: 25%, distance to water: 20%, land cover: 15%, access roads: 10%) and combines them to rank locations. The Analytical Hierarchy Process (AHP) provides a structured method for determining weights.

• Spatial autocorrelation violates classical statistical assumptions
• Hot spot analysis identifies statistically significant spatial clusters
• Optimization models solve complex spatial allocation problems
• MCDA combines multiple criteria into integrated suitability assessments