In an earlier post I provided an overview of the different types of methods that might be used to conduct different types of spatial data analyses.  These included using statistical methods, spatial analysis methods, data science methods and depending on your needs and use of imagery, image analysis methods.

In this post I will go into a bit more detail about six essential spatial analysis function classes that are used to analyse spatial information. You may already know about some of these, none-the-less as technologies and data types continue to evolve so too will the methods needed to process and analyse these data.

Spatial analysis refers to the ability to transform spatial data into different forms and extract meaning as a result of these transformations. A variety of technique are available that use location (where a feature or object is) OR both the location and attribute information (where located, what it is, how much/intensity) about a feature. As such spatial analysis covers a broad range of methods. Today many GIS and other platforms support a range of spatial analysis functions and methods (see GIS ecosystems post).

Broadly, the six function classes include: measurement, network analysis, topological analysis, spatial statistical analysis, surface analysis (including raster-based analysis) and GeoAI as summarized in the figure.  An overview of the functions associated with each class along with a brief description is captured in the table below.

Overview of the six spatial analysis function classes used to analyse spatial information and better understand the world in which we live (Source: GeoSpatialSense, 2024).

Function classFunctionDescription
MeasurementDistance, length, perimeter, area, centroid, buffering, volume, shape, measurement scale conversionOccupied space: Useful for determining the size of a feature/object and how much space something occupies.  

Distances: Useful for calculating straight-line distances between points, distances along paths, arcs or areas.  

Relationships & Interaction: Distance as a measure of separation is a key variable for assessing interactions, defining relationships and connectivity between places and objects.  
Topological analysisAdjacency, polygon overlay, point-in-polygon, line-in-polygon, dissolve, merge, clip, erase, intersect, union, identity, spatial join, and selectionSpatial relationships: Used to describe and analyze spatial relationships among objects and features (e.g. something is next to something else; borders X; is completely within Y, etc.).  
Topological analysis functions can identify features in the landscape that are adjacent or next to each other (contiguous).
Topology is important in modeling connectivity in networks, understanding relationships.  

Think of Venn diagrams from maths class.
Network and location analysisConnectivity, shortest path analysis, routing, service areas, location-allocation modeling, accessibility modelingInvestigates flows through a network. Network is modeled as a set of nodes and the links that connect the nodes.

-Can capture routes, pathways and connections.
– Can be accumulative and/or impedances.  
Think of location-allocation models; supply-chains, service areas, social-network analysis.
Surface analysisSlope, aspect, filtering, line-of-sight, viewsheds, contours, watersheds; surface overlays or multi-criteria decision analysis (MCDA)Representation of the real-world. Useful for defining a spatial concept, model or representation of the real-world. Representing a continuous surface of the real-world and often used to delineate and analyse terrain and other environmental characteristics (e.g. temperature, rainfall, hydrological watersheds, impedances across a landscape (e.g. travel, visibility and signal ranges, etc)  

Neighbourhood and filtering techniques: Filtering techniques include smoothing (remove noise from data to reveal broader trends) and edge enhancement (accentuate contrast and aids in the identification of features).  

Raster-based modelling useful for performing complex mathematical operations that combine and integrate data layers in an array (e.g. Multi-criteria Decision Analysis (MCDA), Fuzzy logic, overlay and weighted overlay methods), etc.
Spatial Statistical analysisSpatial sampling, spatial weights, exploratory data analysis, nearest neighbor analysis, Global and local spatial autocorrelation, spatial interpolation, geostatistics, trend surface analysis. Regression analysis and spatial relationshipsExploratory data analysis: These methods analyze information based on attributes as well as their spatial relationship. Depending on methods these can explore relationships, interactions or comparisons.

Pattern analysis – Descriptive Spatial Statistics, Point pattern Analysis, Hotspot analysis.

Zonal Statistics – descriptive statistics by geographic area
GeoAI – Geospatial Artificial IntelligenceMining of large amounts of data Extraction of information – text, images, features Sensemaking: Grouping and clustering of information Multidimensionality of information (space-time-location); SLAM (simultaneous localization and mapping)  Advanced level techniques that require the integration of spatial/image analysis with data science, statistics, and computer programming to accelerate understanding.

Some of the methods used for GeoAI are already present in some of the software tools that are used to perform different types of spatial analysis.

Image Analysis, Random Forest (e.g. feature identification and extraction) Supervised and unsupervised clustering Natural Language Processing (text extraction, semantic exploration & processing)
Big Data Analysis and mining
Automation of workflows and processes
Data fusion and integration
Spatial-temporal analyses
On-the-fly processing (e.g. SLAM)
Table containing an overview of the spatial analysis classes and functions of GIS. These fall into six classes: measurement, topological analysis, network analysis, surface analysis and statistical analysis and GeoAI.

Sources and references:

Skills for becoming a Spatial Data Scientist, GIScientist, GIS Analyst
getting started: Spatial Data Science, GIScientist…

Sources:

© 2023-present GeoSpatialSense. All Rights Reserved

Unknown's avatar

Published by geospatialsense