Space-Time Hot Spot Analysis
Space-Time Hot Spot Analysis (Emerging Hot Spot Analysis) · Also known as: emerging hot spot analysis, space-time cube hot spot, spatiotemporal hot spot detection, STHA
Space-Time Hot Spot Analysis extends the classic Getis-Ord Gi* statistic across repeated time slices organised in a space-time cube. By testing each location-time bin for statistically significant clustering of high or low values, then examining the sequence of results over time, it identifies whether clusters are new, intensifying, persistent, sporadic, or diminishing — giving analysts a dynamic picture of how hot and cold spots evolve.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use space-time hot spot analysis when you have georeferenced event data collected repeatedly over time (crime incidents, disease cases, traffic accidents, social media posts) and you want to understand not just where clusters occur but how they change. It is appropriate when data can be sensibly aggregated into a regular space-time cube (uniform spatial bins and equal-length time periods) and when the number of time periods is at least 10 to support trend testing. Do not use it when time intervals are irregular or very few (fewer than ~6 periods), when the attribute is a ratio without meaningful denominators for spatial joining, or when the goal is simply to map the current snapshot rather than temporal dynamics — plain Getis-Ord Gi* suffices in that case.
Strengths & limitations
- Reveals temporal dynamics of clustering that a single cross-sectional hot spot map cannot show.
- Produces interpretable, policy-relevant categories (e.g., Emerging vs Persistent hot spots) that guide resource allocation.
- Built on the well-validated Getis-Ord Gi* statistic, so each time-slice result has a rigorous statistical basis.
- Works with any count or rate variable that can be spatially aggregated — crime, disease, pollution, sales, etc.
- Handles large datasets efficiently when implemented in a space-time cube framework.
- Results are sensitive to the chosen spatial bin size and time interval; different aggregations can yield different pattern categories.
- Requires at least 10 time periods for the Mann-Kendall trend test to be meaningful; very short time series may produce unreliable classifications.
- Assumes stationarity of the spatial weight structure; shifting study boundaries or variable neighbourhood definitions complicate interpretation.
- Does not model covariates or control for confounders — it is descriptive, not causal.
- Sparse data in many cells (many zeros) can inflate apparent cold spots and suppress detection of real hot spots.
Frequently asked
How many time periods do I need?
At least 10 time periods are recommended for the Mann-Kendall trend component to produce reliable results. Fewer periods will still generate Gi* values per slice, but the trend classification will be statistically weak.
What is the difference between space-time hot spot analysis and plain Getis-Ord Gi*?
Plain Gi* analyses a single cross-section and identifies where high-value clusters exist right now. Space-time hot spot analysis stacks many Gi* maps over time and adds a trend test, so it can tell you whether a cluster is new, growing, stable, or fading.
How do I choose bin size and time interval?
Bin size should reflect the meaningful geographic scale of the phenomenon (e.g., a city block for crime, a county for disease). Time intervals should match the periodicity of reporting or the expected pace of change. Run a sensitivity analysis with two or three alternatives to check that key findings are robust.
Does the method require a minimum number of events per bin?
There is no strict minimum, but bins with very low counts (near zero) are unlikely to form statistically significant clusters. A commonly applied rule of thumb is to ensure that the average bin count is at least 8–10 before relying on the results.
Can I use continuous values instead of counts?
Yes. The Gi* statistic can be applied to any numeric attribute (counts, rates, temperatures, prices). The space-time cube aggregation step should use the appropriate summary function (sum for counts, mean for continuous measures).
Sources
- Kulldorff, M. (1997). A spatial scan statistic. Communications in Statistics: Theory and Methods, 26(6), 1481–1496. DOI: 10.1080/03610929708831995 ↗
- Emerging Hot Spot Analysis. ArcGIS Pro Documentation. Esri Inc., 2021. link ↗
How to cite this page
ScholarGate. (2026, June 3). Space-Time Hot Spot Analysis (Emerging Hot Spot Analysis). ScholarGate. https://scholargate.app/en/spatial-analysis/space-time-hot-spot-analysis
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Hot Spot AnalysisSpatial analysis↔ compare
- Local Getis-Ord Gi*Spatial analysis↔ compare
- Local Spatial AutocorrelationSpatial analysis↔ compare
- Space-Time Local Indicators of Spatial AssociationSpatial analysis↔ compare
- Space-Time Moran's ISpatial analysis↔ compare