North Carolina statewide field validation data compares three automated detection methods for mobile home parks. The dataset includes field observations from windshield surveys, modeling inputs, and probabilistic confidence scores developed from the evaluation. It was created by Heerae Lee and last updated in June 2026.
Use Cases
- Evaluate computer vision model performance for mobile home park detection based on field validation observations.
- Compare the accuracy of different statewide data sources, such as HIFLD and building footprint datasets.
- Develop classification models to estimate confidence scores for candidate mobile home park locations.
- Support future research on manufactured housing communities using cleaned geospatial datasets.
Strengths
- Field validation observations cover 32 North Carolina counties.
- Evaluates three distinct statewide data sources: a computer vision dataset, the HIFLD dataset, and a building footprint–based dataset.
- Includes reproducible code and metadata documentation alongside cleaned geospatial data.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- Data may reflect geographic bias inherent to the windshield survey methodology across selected counties.
Provenance
- Source
- ODUM Harvested Dataverse
- Collection Method
- Field validation via windshield surveys combined with analysis of three existing geospatial datasets.
- Freshness
- Last updated 2026-06-01 03:10:07; freshness should be verified.
- Geography
- North Carolina, United States