SGVI-2023 v2: Street-Level Seasonal Green View Index

SGVI-2023 v2 systematically fills missing street-level Green View Index values using two optimized predictive models. One model uses Sentinel-1 and Planet imagery, achieving a point-scale correlation coefficient of 0.83084, while another uses Sentinel-2, Planet, and NDVI for improved accuracy with R=0.8789. This process generates a more spatially complete dataset for assessing urban vegetation coverage.

Use Cases

• [Urban vegetation monitoring] based on [street-level Green View Index (GVI) values]
• [Predictive model validation] based on [correlation coefficients and error metrics]
• [Spatial gap analysis] based on [systematically completed missing values]
• [Remote sensing data fusion] based on [Sentinel-1, Sentinel-2, and Planet imagery inputs]

Strengths

• Model performance metrics are explicitly stated, with a correlation coefficient of 0.8789 for the most accurate model.
• The dataset is an updated version designed to address spatial completeness by filling missing GVI values.
• It is publicly available under a CC-BY-4.0 license.

Limitations

• The exact number of rows and specific column names are not provided, limiting precise understanding of dataset structure.
• The dataset's spatial geography and temporal time_range are not explicitly stated.
• The description is duplicated across sources, suggesting potential lack of varied metadata.

Provenance

Source

Yibin Ma

Collection Method

Missing values were filled using two optimized predictive models developed with Sentinel and Planet remote sensing imagery.

Freshness

2026-03-20 03:19:23