Overview
This lecture introduces hotspot analysis in spatial statistics, focusing on selecting the appropriate conceptualization of spatial relationships for analyzing childhood obesity rates in Los Angeles County elementary schools.
Problem Definition and Data
- The analysis investigates spatial clusters ("hotspots") of overweight fifth graders in Los Angeles County schools.
- Data consists of school zones (polygons), each representing aggregated block groups closest to a school, with associated childhood obesity rates.
- Initial visual examination of thematic maps can be subjective and influenced by map classification schemes.
Purpose of Hotspot Analysis
- Hotspot analysis tests whether observed clusters of high or low obesity rates are statistically significant.
- The GI* (Getis-Ord Gi*) statistical method is used for this analysis.
Hotspot Analysis Workflow
- Input feature class: polygons representing school zones with obesity rate data.
- Input field: obesity rates for fifth graders (numeric field to be analyzed).
- Output feature class: location to save results specified in a geodatabase or folder.
Choosing a Conceptualization of Spatial Relationships
- Conceptualization defines how spatial neighbors and influences are determined.
- Fixed Distance Band: neighbors are features within a set distance; provides consistent scale and is often recommended.
- Polygon Contiguity: neighbors are polygons sharing a boundary; suitable when polygons are similar in size.
- Inverse Distance: not recommended for hotspot analysis as it can produce small, isolated hotspots.
- Zone of Indifference: similar to fixed distance band, but features outside the distance are weighted by inverse distance (creates a "fuzzy" boundary).
Recommended Options for Analysis
- Fixed distance band or zone of indifference are suitable choices for this dataset due to variance in polygon sizes in LA County.
- The lecture proceeds with the zone of indifference method.
Key Terms & Definitions
- Hotspot Analysis — Identifying statistically significant spatial clusters of high or low values.
- GI (Getis-Ord Gi)** — A statistical method for determining hotspot significance.
- Conceptualization of Spatial Relationships — The method used to define spatial neighbors in an analysis.
- Fixed Distance Band — Includes all features within a set distance as neighbors.
- Polygon Contiguity — Defines neighbors as adjoining polygons.
- Inverse Distance — Weights neighbors by how close they are; not ideal for hotspot analysis.
- Zone of Indifference — Combination of fixed distance and inverse distance methods for defining neighbors.
Action Items / Next Steps
- Watch part two to learn how to choose an appropriate distance band for the analysis.