Spatial Analysis and Modeling
Spatial Analysis and Modeling is an exciting area under GIScience, which aims to process, analyze, visualize, model, and better understand spatial processes (often with a temporal dimension) that take place on the earth. It includes a set of closely related subareas: agent-based modeling, data analytics and geographic knowledge discovery, social network analysis, spatial and spatiotemporal statistics, and spatial and spatiotemporal modeling / simulation, which are introduced below.
Agent-Based Modeling: Agent-based modeling (ABM), or individual based modeling (IBM) is rooted on the fundamental philosophy of methodological individualism. ABM focuses on the uniqueness of individuals and interactions among them or between these individuals and the associated environment(s), generating patterns or results not analytically tractable from system components and their attributes alone. ABM is thus powerful in explaining or envisioning some complexity features such as emergence, path-dependence, multifinality, and equifinality.
Data Analytics and Geographic Knowledge Discovery: There is an increasing demand in decision makings to extract hidden patterns, trends, and useful information and knowledge from massive high-dimensional spatial datasets. This research area involves various computational, statistical, and geographical methods and techniques (e.g., spatial data mining, machine learning, artificial intelligence) to achieve tasks such as classification, clustering, association, deviation and anomaly detection, trends analysis, generalization, and prediction about spatial phenomena.
Social Network Analysis: A social network is a communication structure of social connections, personal relationships, or group interactions within a local or global community. It is an important interdisciplinary research domain in social science, communication theories, computer science, and mathematics. The study of social networks can be traced back to the famous epidemiological study by Dr. John Snow (1855), who used the locations of cholera death records to figure out the spatial pattern of disease spread and the connection to the contaminated water pump stations in London in the 1800s. The spatialization of social networks can provide valuable information and knowledge discovery for understanding human activities, disease outbreaks, and public opinions.
Spatial and Spatiotemporal Statistics: This area of research aims to depict and understand spatial patterns, such as spatial randomness, lustering, or dispersing of point or area data. Increasingly we turn to add a temporal dimension, generating measures of temporal patterns and dynamics. The so-called space-time analysis or spatiotemporal analysis has arisen to capture and envision variability in both space and time, or in coupled space-time domain.
Spatial and Spatiotemporal Modeling: This area of research focuses on understanding, describing, and predicting spatial or spatiotemporal (also termed as space-time) patterns of a certain phenomenon or variable of interest. Key themes include modeling of wildlife behavior and habitat, land use and land cover change, and human movement or residence patterns. Key techniques include land change modeling, latent trajectory modeling (with eigenvector spatial filtering), agent-based modeling, cellular automata, space-time analysis, and the like.
Watershed Analysis: Watershed analysis and modelling is the science of understanding how water moves through watersheds, and representing that movement in spatial, statistical, and computer models. We can use those models to predict hazards (floods, droughts), forecast water availability and test the results of future water management scenarios. We can predict the movement of pollutants and sediment through the watershed. Watershed analysis uses extensive spatial information, often using remote sensing and GIS layers, such as data on land use, soil type, geology and human activities in the watershed. Water movement is highly variable in space and time, from small scales (e.g. water draining through soil cracks) to large scales (e.g. interactions of surface and groundwater focused within river margins). In watershed analysis we study how to preserve the world’s valuable water resources and use them efficiently.
A Spatiotemporal Approach to Sensitivity Analysis in Human-Environment Systems Models
The integrated uncertainty and sensitivity analysis methods developed in this project allow calculating model output uncertainty and sensitivity and visualizing on maps uncertainty and sensitivity measures. This approach is well suited to spatial models, which deliver modeling results in various map formats. Many spatial models use large input datasets and produce model results comprised of large data sets. Such models require significant computational resources to run model calculations. Similarly, large computational resources are required to run uncertainty and sensitivity analysis for such models. A major contribution of this project is the development of method that makes it possible to reduce the number of calculations necessary to arrive at good approximation of model output uncertainty and sensitive model input factors.
We have organized a symposium under the goal of transforming the science, technology, and application of agent-based models (ABMs, or agent-based modeling—ABM) in the context of social, human-environment, and life sciences. The symposium will be April 20-22, 2017 at San Diego, California. We have assembled a science committee of about 35 exceptional ABM modelers and users, and are accepting participants from various disciplines. The symposium has the following objectives: 1) summarizing the status-quo of ABM, and particularly pinpointing the strengths and weaknesses of ABM; 2) having in-depth discussions centered around a set of topical subareas (to be identified, but likely including model validation, modeling of human decisions, model transparency and reusability, and developing big data friendly ABM); and 3) identifying unique resources, areas of collaboration, impending tasks, and future directions for the ABM community.
- GEOG 385: Spatial Data Analysis
- GEOG 484: Geographic Information Systems
- GEOG 506: Landscape Ecology
- GEOG 576: Advanced Watershed Analysis
- GEOG 584: Geographic Information Systems Applications
- GEOG 585: Quantitative Methods in Geographic Research
- GEOG 589: GIS-Based Decision Support Methods
- GEOG 594: Big Data Science and Analytics Platforms
- GEOG 683: Advanced Geographic Information Systems
- GEOG 683L: Geographic Information Systems Laboratory
- GEOG 780: Seminar in Techniques of Spatial Analysis
Analysis of spatially distributed data including computer applications. Spatial sampling, descriptive statistics for areal data, inferential statistics, use of maps in data analysis.
Procedures for encoding, storage, management, and display of spatial data; theory of computer-assisted map analysis; examination of important geographic information systems.
Links between landscape patterns and ecological processes at a variety of spatial scales to include causes and measures of landscape patterns, effects of landscape patterns on organisms, landscape models, landscape planning and management.
Theory and techniques in watershed analysis. Use of GIS and statistical programming for analyses of geomorphology, hydrology, and water quality data.
Spatial analysis methods in GIS, to include terrain, raster, and network analysis. Feature distributions and patterns. GIS data processing techniques to include spatial interpolation, geocoding, and dynamic segmentation. Designing and executing analytical procedures.
Application of statistical techniques to geographic research to include simple regression and correlation, multiple regression, geographically weighted regression, classification, factor analysis, and computer applications.
Integration of Geographic Information Systems (GIS) with discrete and continuous multiple criteria decision making (MCDM) methods. Applications of MCDM in land use planning, site selection, and resource management spatial decision problems.
Big data science to include analysis, data collection, filtering, GIS, machine learning, processing, text analysis, and visualization. Computational platforms, skills, and tools for conducting big data analytics with real world case studies and examples.
Scripting techniques with Python for automating geoprocessing tasks and developing GIS tools. Use of Bayes’ Theorem in spatial modeling.
Geoprocessing Python scripting techniques with applications to spatial modeling and analysis.
Spatial analytic techniques from image processing, remote sensing, geographic information systems, cartography or quantitative methods. May be repeated with new content. See Class Schedule for specific content.