Skip to contents

Risk Mapping

Source: vignettes/risk-mapping.Rmd
risk-mapping.Rmd

Overview

Mapping is often the final communication stage of a spatial risk workflow.

In riskworkflowr, mapping functions are intended to support:

  • exploratory analysis
  • operational review
  • comparative spatial interpretation
  • communication of broad spatial patterns

The package currently focuses on choropleth-style thematic mapping workflows.

Core workflow

Typical workflows involve:

point events
→ spatial assignment
→ aggregation/counts
→ risk metrics
→ mapping

The main mapping helper function is:

Supported map types

Current map types include:

count
rate
probability
smr
category

These support both continuous and categorical spatial outputs.

Example workflow 

map_risk_choropleth(
  data = analysis_units,
  fill_col = "smr",
  map_type = "smr"
)

Choropleth mapping considerations

Choropleth maps can be useful for communicating broad spatial patterns, but they also introduce important interpretive challenges.

Map outputs are influenced by:

  • classification method
  • colour palette
  • spatial scale
  • denominator quality
  • spatial unit design
  • event rarity
  • data smoothing choices

The same underlying data may appear substantially different depending on these decisions.

Classification methods

Different classification approaches may produce different visual interpretations.

Common approaches include:

  • quantile classification
  • equal interval classification
  • Jenks natural breaks
  • custom thresholds
  • binary or categorical thresholds

No single classification method is universally correct.

The most appropriate approach depends on:

  • the analytical objective
  • data distribution
  • intended audience
  • communication requirements

Rates versus counts

Counts and rates communicate different information.

Counts

Count maps may highlight:

  • total burden
  • operational workload
  • absolute event volume

Rates

Rate maps may highlight:

  • relative concentration
  • comparative occurrence
  • elevated occurrence relative to exposure

Both may be useful, depending on the question being asked.

SMR-style mapping

SMR-style maps may support exploratory identification of areas with comparatively elevated or reduced observed-versus-expected outcomes.

However, these maps should not be interpreted as:

  • causal evidence
  • definitive hotspot detection
  • proof of operational failure
  • precise estimates of individual risk

Interpretation should consider:

  • denominator quality
  • event rarity
  • small number instability
  • broader contextual information

Category mapping

Categorical maps may support communication of:

  • dominant risk types
  • comparative categories
  • grouped classifications
  • future distinctive category workflows

Future versions of riskworkflowr are planned to support grouped category dominance outputs derived from grouped comparative SMR workflows.

H3 and hexagonal mapping

The same mapping workflow may be applied to:

  • administrative boundaries
  • custom polygons
  • H3 grids
  • hexagonal grids

Different spatial unit systems may produce different visual impressions of the same underlying events.

Analysts should carefully consider:

  • resolution
  • sparsity
  • interpretability
  • operational meaning

when mapping fine-scale grids.

Assumptions

Mapping workflows assume:

  • geometry is valid for the processing context
  • spatial units are meaningful
  • denominators are appropriate
  • classification choices are defensible
  • outputs are interpreted carefully

Limitations and pitfalls

Potential issues include:

  • misleading classification
  • unstable rates
  • sparse denominators
  • false precision
  • ecological interpretation risks
  • overinterpretation of visual patterns
  • boundary effects
  • inconsistent spatial scales

Maps should generally be interpreted as exploratory communication tools rather than definitive analytical conclusions.

Alternative approaches

Alternative visualisation approaches may include:

  • proportional symbols
  • density surfaces
  • kernel density estimation
  • interactive web mapping
  • uncertainty mapping
  • temporal animation
  • small multiple comparisons
  • bivariate mapping

The most appropriate method depends on the intended analytical and communication objective.

Potential future directions

The following areas are currently being explored or considered as possible future enhancements to riskworkflowr.

These items are exploratory only and should not be interpreted as guaranteed future functionality.

Potential areas include:

  • grouped/category SMR workflows
  • distinctive category identification
  • richer H3 workflows
  • improved uncertainty handling
  • enhanced mapping support