Spatial Models and Geometric Representation: How Mathematical Models Represent Place

Last Updated June 12, 2026

Spatial models and geometric representation describe systems where location, shape, distance, direction, scale, region, boundary, and spatial arrangement matter. Instead of treating variables as detached from place, spatial models ask how phenomena are distributed across space and how geometry shapes relationships.

A spatial model is useful when place is not merely background context but part of the model’s structure. A disease outbreak, transportation network, watershed, habitat corridor, floodplain, supply route, service area, urban neighborhood, or climate-risk zone cannot be fully understood without representing where things are and how spatial relationships affect them.

Geometric representation gives spatial reasoning mathematical form. Points, lines, polygons, grids, surfaces, fields, coordinate systems, distance metrics, buffers, adjacency, topology, projections, and spatial transformations allow modelers to translate location and form into computational objects.

Spatial modeling is powerful because it makes geometry part of explanation. A model can ask whether proximity changes exposure, whether distance changes access, whether boundaries shape risk, whether clusters reveal structure, whether routes constrain movement, whether resolution alters conclusions, and whether spatial patterns support or undermine a decision.

Why Spatial Models Matter

Spatial models matter because many systems are shaped by where things are. Distance affects travel time. Boundaries affect jurisdiction. Elevation affects flooding. Adjacency affects spread. Density affects exposure. Connectivity affects access. Shape affects flow. Spatial arrangement can change outcomes even when the same variables are present.

Without spatial representation, a model may incorrectly treat separated things as equally accessible, clustered risks as independent, boundaries as irrelevant, or averaged areas as internally uniform. Spatial modeling helps prevent place-based structure from disappearing into aggregate numbers.

A spatial model does not merely add a map to a result. It makes spatial relationships part of the model’s structure, computation, validation, and interpretation.

Back to top ↑

What a Spatial Model Is

A spatial model represents phenomena using location, geometry, distance, adjacency, surface, region, field, or spatial interaction. It may be descriptive, explanatory, predictive, or decision-supportive.

The spatial component \(s\) may be a coordinate pair, polygon, grid cell, network location, distance matrix, neighborhood set, surface value, or geometric relation. The model’s meaning depends on how that spatial component is defined.

Spatial models require both mathematical and geographic judgment. The same data can produce different conclusions depending on scale, projection, boundary, distance metric, neighborhood definition, and aggregation level.

Back to top ↑

Geometry as Representation

Geometry is not neutral decoration. It is a modeling choice. A real-world object can be represented as a point, line, polygon, grid cell, surface, network node, or region. Each representation preserves some information and discards other information.

A hospital may be modeled as a point for access analysis, a polygon for land-use analysis, a node in an infrastructure network, or a service region for planning. None of these is simply “the hospital.” Each is a representation fitted to a purpose.

Good spatial modeling begins by asking what geometry should mean. The representation must match the model purpose, data quality, scale, and decision context.

Back to top ↑

Points, Lines, Polygons, Grids, and Surfaces

Spatial models often rely on a small set of geometric primitives. These primitives are simple, but their meaning varies by context.

Lines represent paths, edges, rivers, roads, boundaries, routes, or transects. Polygons represent regions, parcels, zones, watersheds, districts, service areas, or habitats. Grids divide space into cells. Surfaces represent continuous variation such as elevation, temperature, pollution, or risk.

The same phenomenon may require multiple geometries. A flood model may combine elevation surfaces, river lines, watershed polygons, infrastructure points, and road networks. Spatial modeling often requires layered representation rather than one geometry alone.

Back to top ↑

Coordinate Systems, Projections, and Reference Frames

Spatial models require coordinate systems. Coordinates define how locations are represented numerically. A coordinate pair has no complete meaning unless the reference system is known.

Projected coordinate systems represent curved Earth locations on a flat plane. Every projection distorts something: area, shape, direction, distance, or scale. The relevant question is not whether distortion exists, but whether the chosen projection is appropriate for the model’s purpose.

Spatial model validation should include reference-system review. If distances, areas, buffers, or overlays are computed in the wrong coordinate system, the model may be numerically precise but spatially wrong.

Back to top ↑

Distance, Direction, Buffers, and Accessibility

Distance is one of the most common spatial model components. But distance is not one thing. Euclidean distance, network distance, travel time, cost distance, perceived distance, and barrier-adjusted distance may all produce different conclusions.

Accessibility models often ask whether people, resources, risks, or services are reachable within a distance, travel-time, cost, or effort threshold.

Distance measures should match the phenomenon. A straight-line buffer may be reasonable for some exposure questions, but misleading for transportation access if rivers, highways, transit routes, or elevation barriers matter.

Back to top ↑

Adjacency, Topology, and Neighborhood Structure

Topology concerns spatial relationships that remain meaningful even when exact distances or shapes change. Adjacency, containment, intersection, connectivity, and overlap are topological relationships.

In many spatial models, relationships are defined by neighborhoods: which regions touch, which grid cells are adjacent, which parcels share boundaries, which locations fall inside a zone, or which network nodes are connected.

Neighborhood definitions are modeling assumptions. Contiguity, distance bands, nearest neighbors, network links, administrative boundaries, and ecological connectivity can all define different spatial relationships.

Back to top ↑

Spatial Fields, Surfaces, and Interpolation

Some spatial models represent phenomena as fields: values that vary continuously or semi-continuously across space. Temperature, elevation, pollution concentration, rainfall, risk, population density, and habitat suitability can be modeled as spatial fields or surfaces.

When observations exist only at sampled locations, interpolation estimates values at unsampled locations. Interpolation methods embed assumptions about smoothness, distance decay, spatial autocorrelation, and uncertainty.

Spatial fields can make patterns visible, but they can also create an illusion of continuous knowledge. Interpolated surfaces should be accompanied by uncertainty, sampling design review, and sensitivity to method choices.

Back to top ↑

Scale, Resolution, and Aggregation

Spatial conclusions depend on scale and resolution. A pattern visible at one scale may disappear, reverse, or fragment at another. Aggregating observations into regions can change apparent relationships because boundaries and units influence the statistics.

This is one reason spatial modeling requires special care. A neighborhood-level model, block-level model, grid-cell model, and household-level model may tell different stories even when based on related data.

Scale is not a technical afterthought. It is part of model design. Spatial models should document resolution, extent, aggregation, and boundary logic before outputs are interpreted.

Back to top ↑

Spatial Interaction, Movement, and Diffusion

Spatial models often represent interaction across locations. People travel. Species migrate. Goods move. Pollution spreads. Fire propagates. Information diffuses. Patients seek care. Workers commute. Water flows. These processes depend on distance, direction, barriers, networks, gradients, and local conditions.

Spatial diffusion models describe how a quantity spreads through space. Network models describe movement along links. Agent-based models can simulate individual movement through spatial environments. These approaches can be combined when spatial interaction is central.

Spatial interaction models must distinguish geometric possibility from actual behavior. A short distance may not imply movement if barriers, cost, law, safety, habit, capacity, or institutions prevent access.

Back to top ↑

Geometric Transformations and Model Operations

Spatial models use geometric operations to transform, combine, compare, and derive spatial objects. These operations can be mathematically precise but still conceptually risky if the underlying representation is weak.

Common operations include buffering, overlay, union, intersection, clipping, dissolving, simplification, reprojection, rasterization, vectorization, interpolation, and distance calculation.

Every spatial operation should be traceable. A reproducible spatial model documents which operations were performed, why they were selected, and how sensitive results are to operational choices.

Back to top ↑

Mathematical Lens: Space as Model Structure

Spatial modeling can be formalized by treating locations as elements of a spatial domain. Let \(D\) be a spatial domain and \(s\in D\) a location. A spatial field assigns a value to each location:

Distance is a function over pairs of locations:

Neighborhoods can be defined by a distance threshold:

Spatial interaction can be represented through a spatial weight matrix:

Geometric transformation can be represented as a function from one coordinate representation to another:

This mathematical lens shows that spatial models are not simply visual models. They are formal structures involving domains, functions, metrics, neighborhoods, transformations, surfaces, and relations.

Back to top ↑

Example: Service Access and Exposure Model

Consider a public service access model. The goal is to estimate which neighborhoods have reasonable access to emergency clinics and which populations fall within a heat-risk exposure zone.

The model uses points for clinics, polygons for neighborhoods, a surface for heat-risk intensity, and distance or travel-time measures for accessibility. It may also include population counts, age structure, transit availability, and neighborhood-level vulnerability indicators.

A simple accessibility score might combine clinic capacity with travel cost. A simple exposure score might overlay population with heat-risk intensity. A decision-support version might identify neighborhoods with high exposure and low service access.

This example illustrates the importance of spatial assumptions. Results may change if access is measured by straight-line distance instead of travel time, if neighborhoods are aggregated differently, if the heat-risk surface is low resolution, or if clinic capacity is ignored.

Spatial models can support better decisions, but only when their geometric representations, distance measures, and uncertainty are visible to the people using the results.

Back to top ↑

Calibration, Validation, Sensitivity, and Spatial Uncertainty

Spatial models require validation because location data, geometry, projection, scale, and spatial assumptions can all introduce error. A visually polished map can hide uncertainty, misalignment, missing data, boundary problems, or inappropriate distance measures.

Spatial uncertainty should be communicated directly. This includes positional uncertainty, boundary uncertainty, measurement uncertainty, interpolation uncertainty, aggregation uncertainty, and model-form uncertainty.

Back to top ↑

Ethical Stakes of Spatial Modeling

Spatial models can shape decisions about policing, zoning, health services, disaster response, housing, infrastructure, environmental regulation, conservation, taxation, surveillance, and public investment. They can make inequality visible, but they can also reproduce or intensify it.

Responsible spatial modeling treats maps and geometric outputs as decision-support tools, not automatic authority. The model should clarify how space matters while making uncertainty, assumptions, and social consequences visible.

Back to top ↑

Python Workflow: Spatial Register and Accessibility Diagnostics

The Python workflow below uses dependency-light spatial calculations to create a spatial model register, compute point-to-service distances, summarize accessibility, and export an audit card. It avoids specialized GIS dependencies so the core modeling logic remains transparent.

# spatial_models_geometric_representation_workflow.py
# Dependency-light workflow for spatial registers and accessibility diagnostics.

from __future__ import annotations

from dataclasses import asdict, dataclass
from pathlib import Path
import csv
import json
import math


ARTICLE_ROOT = Path(__file__).resolve().parents[1]
OUTPUTS = ARTICLE_ROOT / "outputs"
TABLES = OUTPUTS / "tables"
JSON_DIR = OUTPUTS / "json"


@dataclass(frozen=True)
class SpatialRecord:
    key: str
    component_type: str
    geometry_or_structure: str
    interpretation: str
    review_question: str
    status: str


@dataclass(frozen=True)
class Location:
    key: str
    kind: str
    x: float
    y: float
    value: float


def spatial_register():
    return [
        SpatialRecord(
            key="point_geometry",
            component_type="geometry",
            geometry_or_structure="p=(x,y)",
            interpretation="Facilities and observations are represented as point coordinates.",
            review_question="Does point geometry oversimplify area, shape, or access?",
            status="review",
        ),
        SpatialRecord(
            key="euclidean_distance",
            component_type="distance_metric",
            geometry_or_structure="sqrt((x_i-x_j)^2+(y_i-y_j)^2)",
            interpretation="Straight-line distance is used as a transparent baseline.",
            review_question="Should network distance or travel time replace straight-line distance?",
            status="review",
        ),
        SpatialRecord(
            key="service_access",
            component_type="accessibility_metric",
            geometry_or_structure="capacity / (1 + distance)",
            interpretation="Service capacity is discounted by distance.",
            review_question="Does the accessibility metric match the decision context?",
            status="review",
        ),
        SpatialRecord(
            key="spatial_uncertainty",
            component_type="validation_diagnostic",
            geometry_or_structure="distance and boundary sensitivity",
            interpretation="Spatial results require sensitivity checks.",
            review_question="Are conclusions robust to distance and scale assumptions?",
            status="active",
        ),
    ]


def sample_locations():
    return [
        Location("neighborhood_a", "demand", 0.0, 0.0, 1200.0),
        Location("neighborhood_b", "demand", 2.0, 1.0, 900.0),
        Location("neighborhood_c", "demand", 4.0, 2.5, 1400.0),
        Location("neighborhood_d", "demand", 6.0, 1.5, 700.0),
        Location("clinic_1", "service", 1.0, 0.5, 500.0),
        Location("clinic_2", "service", 5.5, 2.0, 650.0),
        Location("clinic_3", "service", 3.0, 4.0, 400.0),
    ]


def euclidean_distance(a, b):
    return math.sqrt((a.x - b.x) ** 2 + (a.y - b.y) ** 2)


def accessibility_rows(locations):
    demand = [item for item in locations if item.kind == "demand"]
    services = [item for item in locations if item.kind == "service"]

    rows = []
    for area in demand:
        nearest_service = min(services, key=lambda service: euclidean_distance(area, service))
        nearest_distance = euclidean_distance(area, nearest_service)
        accessibility = sum(service.value / (1.0 + euclidean_distance(area, service)) for service in services)
        exposure_score = area.value / (1.0 + accessibility)

        rows.append({
            "demand_location": area.key,
            "population_or_demand": round(area.value, 4),
            "nearest_service": nearest_service.key,
            "nearest_distance": round(nearest_distance, 6),
            "accessibility_score": round(accessibility, 6),
            "low_access_exposure_score": round(exposure_score, 6),
        })

    return rows


def spatial_risk_score(record):
    score = {"active": 1.0, "review": 5.0, "revise": 8.0, "archive": 2.0}.get(
        record.status.lower(),
        4.0,
    )
    text = f"{record.component_type} {record.geometry_or_structure} {record.review_question}".lower()
    for term in ["distance", "geometry", "access", "boundary", "scale", "uncertainty"]:
        if term in text:
            score += 1.0
    return round(score, 3)


def write_csv(path, rows):
    path.parent.mkdir(parents=True, exist_ok=True)
    if not rows:
        raise ValueError(f"No rows supplied for {path}")
    with path.open("w", newline="", encoding="utf-8") as handle:
        writer = csv.DictWriter(handle, fieldnames=list(rows[0].keys()))
        writer.writeheader()
        writer.writerows(rows)


def write_json(path, payload):
    path.parent.mkdir(parents=True, exist_ok=True)
    with path.open("w", encoding="utf-8") as handle:
        json.dump(payload, handle, indent=2, sort_keys=True)


def main():
    records = spatial_register()
    locations = sample_locations()
    access = accessibility_rows(locations)

    register_rows = [
        {**asdict(record), "spatial_risk_score": spatial_risk_score(record)}
        for record in records
    ]

    write_csv(TABLES / "spatial_model_register.csv", register_rows)
    write_csv(TABLES / "spatial_locations.csv", [asdict(item) for item in locations])
    write_csv(TABLES / "spatial_accessibility_diagnostics.csv", access)

    write_json(JSON_DIR / "spatial_model_audit_card.json", {
        "article": "Spatial Models and Geometric Representation",
        "records": register_rows,
        "accessibility_diagnostics": access,
        "audit_checks": [
            "coordinate system is documented",
            "geometry type matches model purpose",
            "distance metric is justified",
            "scale and aggregation are reviewed",
            "spatial uncertainty is communicated",
        ],
    })

    print("Spatial models and geometric representation workflow complete.")
    print(f"Wrote outputs to {OUTPUTS}")


if __name__ == "__main__":
    main()

This workflow treats spatial choices as auditable modeling decisions. It records geometry, distance metric, accessibility logic, validation questions, and uncertainty review rather than presenting spatial results as self-explanatory map outputs.

Back to top ↑

R Workflow: Spatial Review and Distance Diagnostics

The R workflow below reviews generated spatial outputs, classifies low-access demand areas, and creates a base R plot of accessibility scores by location.

# spatial_models_geometric_representation_review.R
# Base R workflow for spatial review and distance diagnostics.

args <- commandArgs(trailingOnly = FALSE)
file_arg <- grep("^--file=", args, value = TRUE)

if (length(file_arg) > 0) {
  script_path <- normalizePath(sub("^--file=", "", file_arg[1]), mustWork = TRUE)
  article_root <- normalizePath(file.path(dirname(script_path), ".."), mustWork = TRUE)
} else {
  article_root <- getwd()
}

tables_dir <- file.path(article_root, "outputs", "tables")
figures_dir <- file.path(article_root, "outputs", "figures")
dir.create(tables_dir, recursive = TRUE, showWarnings = FALSE)
dir.create(figures_dir, recursive = TRUE, showWarnings = FALSE)

access_path <- file.path(tables_dir, "spatial_accessibility_diagnostics.csv")
register_path <- file.path(tables_dir, "spatial_model_register.csv")

if (!file.exists(access_path)) {
  stop("Missing spatial accessibility outputs. Run the Python workflow first.")
}

access <- read.csv(access_path, stringsAsFactors = FALSE)
access$accessibility_score <- as.numeric(access$accessibility_score)
access$nearest_distance <- as.numeric(access$nearest_distance)
access$low_access_exposure_score <- as.numeric(access$low_access_exposure_score)

access$review_class <- ifelse(
  access$accessibility_score < median(access$accessibility_score, na.rm = TRUE),
  "lower access review",
  "routine review"
)

write.csv(
  access,
  file.path(tables_dir, "r_spatial_accessibility_review_summary.csv"),
  row.names = FALSE
)

if (file.exists(register_path)) {
  register <- read.csv(register_path, stringsAsFactors = FALSE)

  register$priority <- ifelse(
    register$spatial_risk_score >= 8,
    "high",
    ifelse(register$spatial_risk_score >= 6, "medium", "low")
  )

  write.csv(
    register,
    file.path(tables_dir, "r_spatial_model_review_queue.csv"),
    row.names = FALSE
  )
}

png(file.path(figures_dir, "r_spatial_accessibility_scores.png"), width = 1100, height = 720)

scores <- access$accessibility_score
names(scores) <- access$demand_location

if (length(scores) > 0 && any(is.finite(scores))) {
  barplot(
    sort(scores, decreasing = TRUE),
    las = 2,
    ylab = "Accessibility score",
    main = "Spatial Accessibility Diagnostics"
  )
  grid()
} else {
  plot.new()
  title(main = "Spatial Accessibility Diagnostics")
  text(0.5, 0.5, "No finite accessibility values available.")
}

dev.off()

print(access)

The R layer supports review by identifying locations with lower accessibility and by generating outputs that can be compared across alternative distance metrics, service capacities, or boundary assumptions.

Back to top ↑

Haskell Workflow: Typed Spatial Model Records

Haskell is useful for this article because spatial model components should not collapse into one informal category. Geometry, distance metrics, coordinate systems, accessibility metrics, and validation diagnostics have different meanings.

{-# OPTIONS_GHC -Wall #-}

module Main where

data SpatialComponent
  = GeometryDefinition
  | CoordinateSystem
  | DistanceMetric
  | NeighborhoodRule
  | AccessibilityMetric
  | SpatialField
  | ValidationDiagnostic
  deriving (Eq, Show)

data ReviewStatus
  = Active
  | RequiresReview
  | RequiresValidation
  | RequiresSensitivityTest
  | Revise
  deriving (Eq, Show)

data SpatialRecord = SpatialRecord
  { key :: String
  , component :: SpatialComponent
  , geometryOrStructure :: String
  , interpretation :: String
  , reviewFocus :: String
  , status :: ReviewStatus
  } deriving (Eq, Show)

spatialRegister :: [SpatialRecord]
spatialRegister =
  [ SpatialRecord
      "point_geometry"
      GeometryDefinition
      "p=(x,y)"
      "Facilities and observations are represented as point coordinates."
      "Geometry simplification."
      RequiresReview
  , SpatialRecord
      "euclidean_distance"
      DistanceMetric
      "sqrt((x_i-x_j)^2+(y_i-y_j)^2)"
      "Straight-line distance is used as a transparent baseline."
      "Distance validity."
      RequiresReview
  , SpatialRecord
      "service_access"
      AccessibilityMetric
      "capacity / (1 + distance)"
      "Service capacity is discounted by distance."
      "Decision relevance."
      RequiresValidation
  , SpatialRecord
      "spatial_uncertainty"
      ValidationDiagnostic
      "distance and boundary sensitivity"
      "Spatial results require sensitivity checks."
      "Uncertainty communication."
      Active
  ]

needsReview :: SpatialRecord -> Bool
needsReview item =
  case status item of
    Active -> False
    _ -> True

main :: IO ()
main = do
  putStrLn "Typed spatial model records:"
  mapM_ print spatialRegister

  putStrLn "\nSpatial records requiring review:"
  mapM_ print (filter needsReview spatialRegister)

This typed layer supports spatial model governance by making geometry, metrics, and validation roles explicit before map-like outputs are interpreted as evidence.

Back to top ↑

GitHub Repository

The companion repository for this article is designed as a reproducible mathematical-modeling workspace. It contains article-specific code, data, documentation, notebooks, schemas, and generated outputs for spatial model registers, geometric representation, distance diagnostics, accessibility review, typed Haskell spatial records, validation planning, and reproducible computational workflows.

Back to top ↑

A Practical Method for Spatial Model Design

Spatial model design should begin with the spatial question. The model should clarify whether it is concerned with distance, access, exposure, clustering, movement, adjacency, boundaries, surfaces, or geometric form.

This method helps prevent spatial modeling from becoming map decoration. It connects geometric representation to purpose, computation, validation, uncertainty, and responsible interpretation.

Back to top ↑

Common Pitfalls

Spatial models can fail even when the map looks convincing. Many failures arise from projection errors, inappropriate distance metrics, aggregation bias, unclear boundaries, or overinterpretation of visual patterns.

• Map realism error: assuming a model is accurate because it looks geographically detailed.
• Wrong distance metric: using straight-line distance when travel time, network cost, or barriers matter.
• Projection neglect: computing distance or area in an unsuitable coordinate system.
• Boundary dependence: treating administrative boundaries as natural model units without review.
• Scale blindness: ignoring how resolution and aggregation change conclusions.
• Overlay overconfidence: treating layer intersection as precise when data are uncertain or misaligned.
• Surface illusion: presenting interpolated surfaces as if every location were directly observed.
• Cluster overinterpretation: seeing meaningful clusters where bandwidth, scale, or sampling bias created them.
• Privacy neglect: exposing sensitive locations or fine-grained population patterns.
• Decision overreach: using exploratory spatial outputs as if they were validated allocation rules.

These pitfalls can be reduced through coordinate-system checks, geometry validation, distance-metric comparison, scale sensitivity, boundary review, uncertainty communication, and careful separation between visual display and model evidence.

Back to top ↑

Conclusion: Geometry Makes Place Modelable

Spatial models and geometric representation make place modelable. They allow mathematical modeling to represent not only variables and relationships, but also location, distance, shape, adjacency, surface, boundary, movement, exposure, and access.

Their strength is spatial clarity. A spatial model can show how outcomes depend on where things are, how far apart they are, what boundaries contain them, what routes connect them, what surfaces influence them, and what scale reveals or hides.

But spatial models require discipline. Geometry must match purpose. Coordinate systems must be documented. Distance metrics must be justified. Boundaries and scale must be reviewed. Spatial uncertainty must be communicated. A map-like output should not be treated as proof simply because it is visually persuasive.

Used responsibly, spatial models help analysts reason about complex systems where place matters. They make spatial assumptions visible, testable, and open to review, supporting better decisions in public health, infrastructure, ecology, climate risk, logistics, urban planning, sustainability, and policy.

Back to top ↑

Related Articles

• What Is Mathematical Modeling?
• Model Boundaries, Scale, and Scope
• Dimensional Analysis, Units, and Scale
• State Variables and System Representation
• Network Models and Graph Structures
• Agent-Based Models and Emergent Behavior
• Simulation and Computational Modeling
• Uncertainty in Mathematical Models
• Mathematical Modeling in Ecology and Sustainability
• Mathematical Modeling in Policy and Public Systems

Back to top ↑

Further Reading

• de Smith, M.J., Goodchild, M.F. and Longley, P.A. (2021) Geospatial Analysis: A Comprehensive Guide to Principles, Techniques and Software Tools. 6th edn. Available at: https://www.spatialanalysisonline.com/
• Goodchild, M.F. (1992) ‘Geographical information science’, International Journal of Geographical Information Systems, 6(1), pp. 31–45.
• Longley, P.A., Goodchild, M.F., Maguire, D.J. and Rhind, D.W. (2015) Geographic Information Science and Systems. 4th edn. Hoboken, NJ: Wiley.
• O’Sullivan, D. and Unwin, D.J. (2010) Geographic Information Analysis. 2nd edn. Hoboken, NJ: Wiley.
• Burrough, P.A., McDonnell, R.A. and Lloyd, C.D. (2015) Principles of Geographical Information Systems. 3rd edn. Oxford: Oxford University Press.
• Fotheringham, A.S., Brunsdon, C. and Charlton, M. (2000) Quantitative Geography: Perspectives on Spatial Data Analysis. London: SAGE.
• Cressie, N. (1993) Statistics for Spatial Data. Revised edn. New York: Wiley.
• Haining, R. and Li, G. (2020) Modelling Spatial and Spatial-Temporal Data: A Bayesian Approach. Boca Raton, FL: CRC Press.
• Cliff, A.D. and Ord, J.K. (1981) Spatial Processes: Models and Applications. London: Pion.
• Okabe, A. and Sugihara, K. (2012) Spatial Analysis Along Networks: Statistical and Computational Methods. Chichester: Wiley.

Back to top ↑

References

• Burrough, P.A., McDonnell, R.A. and Lloyd, C.D. (2015) Principles of Geographical Information Systems. 3rd edn. Oxford: Oxford University Press.
• Cliff, A.D. and Ord, J.K. (1981) Spatial Processes: Models and Applications. London: Pion.
• Cressie, N. (1993) Statistics for Spatial Data. Revised edn. New York: Wiley.
• de Smith, M.J., Goodchild, M.F. and Longley, P.A. (2021) Geospatial Analysis: A Comprehensive Guide to Principles, Techniques and Software Tools. 6th edn. Available at: https://www.spatialanalysisonline.com/
• Fotheringham, A.S., Brunsdon, C. and Charlton, M. (2000) Quantitative Geography: Perspectives on Spatial Data Analysis. London: SAGE.
• Goodchild, M.F. (1992) ‘Geographical information science’, International Journal of Geographical Information Systems, 6(1), pp. 31–45.
• Haining, R. and Li, G. (2020) Modelling Spatial and Spatial-Temporal Data: A Bayesian Approach. Boca Raton, FL: CRC Press.
• Longley, P.A., Goodchild, M.F., Maguire, D.J. and Rhind, D.W. (2015) Geographic Information Science and Systems. 4th edn. Hoboken, NJ: Wiley.
• Okabe, A. and Sugihara, K. (2012) Spatial Analysis Along Networks: Statistical and Computational Methods. Chichester: Wiley.
• O’Sullivan, D. and Unwin, D.J. (2010) Geographic Information Analysis. 2nd edn. Hoboken, NJ: Wiley.
• Tobler, W.R. (1970) ‘A computer movie simulating urban growth in the Detroit region’, Economic Geography, 46, pp. 234–240.
• Worboys, M. and Duckham, M. (2004) GIS: A Computing Perspective. 2nd edn. Boca Raton, FL: CRC Press.

Back to top ↑