Spatial Analysis Using R Training Course

Participants can import common spatial formats and inspect their essential properties.

• Read GeoPackage, GeoJSON, shapefile and CSV location data into R
• Import GeoTIFF rasters and inspect bands, resolution and missing values
• Examine feature counts, field types, extents and geometry classes
• Identify invalid coordinates, duplicate records and incomplete location attributes
• Use glimpse, summary and plot for initial spatial data inspection
• Document data provenance, collection dates, licences and known limitations

Participants can select coordinate reference systems and correct common geometry issues.

• Geographic and projected coordinate systems distinguished by measurement purpose
• EPSG codes identify coordinate reference system definitions consistently
• Transform layers with st_transform before distance, area or buffer calculations
• Select local projected systems appropriate for metric analysis requirements
• Validate polygon geometry using st_is_valid and st_make_valid
• Detect axis-order, datum and unit mismatches before combining datasets

Participants can prepare spatial attributes for reliable joins, summaries and modelling.

• Standardise identifiers, date fields, categories and numeric measurement units
• Join tabular indicators to boundaries using checked key relationships
• Aggregate records by administrative area, facility or service catchment
• Calculate rates, densities and denominators before mapping area comparisons
• Handle missing values without converting absence into geographic zeroes
• Create derived variables with documented formulas and source fields

Participants can use spatial predicates and overlays to create analytical datasets.

• Spatial predicates test intersects, within, contains, touches and overlaps relationships
• Clip features with st_intersection using a defined study-area boundary
• Dissolve polygons after grouping by region, class or reporting unit
• Extract point attributes from containing administrative or service polygons
• Filter features by geometry type, attribute condition and spatial extent
• Compare intersection, union, difference and symmetric difference outputs

Participants can measure proximity and define defensible service catchment areas.

• Generate fixed-distance buffers around facilities, incidents and monitoring locations
• Calculate nearest-feature distances using st_distance and spatial indexing
• Summarise population, assets or demand points within buffer zones
• Compare Euclidean distance assumptions with network travel-time requirements
• Build concentric rings for graduated access and exposure analysis
• Report buffer parameters, projected units and boundary treatment explicitly

Participants can prepare point-location data and assess its positional fitness for purpose.

• Geocode addresses only after standardising locality, street and postal fields
• Assess match quality through confidence scores, duplicates and unmatched records
• Check coordinates against expected country, district and operational boundaries
• Identify swapped latitude-longitude values and implausible zero-coordinate records
• Mask sensitive point locations through aggregation, jittering or displacement
• Record positional accuracy and privacy decisions within metadata notes

Participants can produce thematic maps that accurately communicate spatial comparisons.

• Choose choropleth, proportional-symbol, dot-density and reference map types
• Classify choropleths using quantiles, equal intervals and natural breaks
• Apply sequential, diverging and qualitative colour schemes appropriately
• Design legends, labels, scale bars and north arrows purposefully
• Use tmap layouts for consistent map series and reporting outputs
• Avoid misleading visual comparisons caused by area size or classification choices

Participants can prepare raster datasets for surface-based spatial analysis.

• Inspect raster coordinate systems, extent, resolution, layers and cell values
• Crop and mask rasters to defined study-area boundaries
• Resample raster layers using nearest-neighbour or bilinear interpolation methods
• Align grids before cell-by-cell calculations and multi-layer modelling
• Reclassify continuous values into analytical categories and thresholds
• Extract raster values at points, lines and polygon zones

Participants can derive terrain indicators and interpret environmental raster surfaces.

• Derive slope, aspect and hillshade from digital elevation models
• Calculate terrain ruggedness for route, habitat and hazard assessments
• Apply focal operations to smooth, summarise or detect local variation
• Calculate zonal statistics for administrative areas and management units
• Interpret resolution effects on terrain measures and environmental indicators
• Document raster source dates, processing steps and resampling choices

Participants can create and evaluate interpolated surfaces from sampled point data.

• Assess sampling distribution before interpolating rainfall, pollution or soil measurements
• Create inverse distance weighting surfaces with stated power parameters
• Compare ordinary kriging assumptions with inverse distance weighting outputs
• Set interpolation extent, cell size and neighbourhood search parameters
• Validate predictions using holdout points and residual error summaries
• Map prediction uncertainty separately from estimated measurement surfaces

Participants can assess whether observed point events are clustered, dispersed or random.

• Define an observation window before analysing incident or event locations
• Map event intensity using kernel density estimation parameters explicitly
• Compare observed patterns with complete spatial randomness assumptions
• Use nearest-neighbour distances to assess local point concentration
• Interpret edge effects and reporting bias in observed event datasets
• Distinguish event counts from exposure-adjusted rates and risks

Participants can test and interpret spatial dependence in area-based indicators.

• Construct spatial weights from contiguity, distance bands or nearest neighbours
• Calculate global Moran's I for overall spatial autocorrelation assessment
• Map local Moran's I clusters and spatial outlier areas
• Interpret high-high, low-low, high-low and low-high cluster categories
• Assess multiple-testing concerns when interpreting local significance maps
• Separate spatial association findings from causal claims about outcomes

Participants can diagnose when conventional regression requires spatial modelling approaches.

• Test regression residuals for spatial autocorrelation before drawing conclusions
• Compare spatial lag and spatial error model interpretations
• Specify geographically weighted regression as an exploratory local model
• Assess multicollinearity, residual patterns and model fit diagnostics
• Interpret coefficients without confusing association, prediction and causation
• Report neighbourhood definitions and model assumptions alongside results

Participants can build transparent raster-based suitability models for location decisions.

• Frame site-selection criteria, constraints and decision rules with stakeholders
• Reclassify slope, distance, land cover and risk criteria
• Standardise criteria scores before combining layers with different measurement scales
• Apply weighted linear combination for transparent suitability scoring
• Test sensitivity to alternative criterion weights and threshold values
• Exclude restricted areas using Boolean constraint masks before ranking

Participants can distinguish network accessibility measures from straight-line proximity measures.

• Represent road networks as nodes, edges, direction and travel costs
• Calculate shortest paths using distance, time or impedance weights
• Define service areas from travel-time thresholds and network constraints
• Compare facility access results with Euclidean buffer approximations
• Identify underserved populations outside defined travel-time service areas
• State network date, speed assumptions and excluded transport modes

Participants can package spatial analysis outputs for review, reuse and audit.

• Organise scripts into import, preparation, analysis and reporting stages
• Parameterise study areas, dates and thresholds for repeatable updates
• Generate maps, tables and narrative findings using Quarto documents
• Record package versions, sources and processing decisions for audit
• Apply quality checks before publishing figures, files and derived datasets
• Write methods statements that explain uncertainty and analytical limitations

Participants can assess the credibility and decision relevance of spatial findings.

• Translate decision questions into measurable spatial indicators and comparison groups
• Review maps for denominator choice, classification and visual bias
• Challenge causal interpretations based solely on geographic correlation patterns
• Compare alternative assumptions through scenario maps and sensitivity tables
• Prioritise findings by operational consequence, uncertainty and implementation feasibility
• Present recommendations with clear geographic scope and evidence boundaries

Participants can build, review and present a complete reproducible spatial analysis addressing a defined decision problem.

• Define a decision question, study area, indicators and analytical boundaries
• Prepare source data, validate geometries and document coordinate reference systems
• Select vector, raster or statistical methods suited to evidence needs
• Build reproducible R scripts, maps, tables and quality-control checks
• Review assumptions, uncertainty, privacy constraints and interpretation limitations
• Present findings, recommendations and implementation implications to a decision audience