Analysis of Urban Green Spaces Using Support Vector Machine in Urban West Region of Zanzibar

Integrating remote sensing techniques with Machine learning-based methods is crucial for analyzing land spatial structures. This study employs the Support Vector Machine to analyze Urban Green Space in the Urban West Region of Zanzibar. The analysis focused on evaluating Support Vector Machine performance in remote sensing imagery classification, assessing green landscape connectivity, and examining geospatial green change trends over 10 10-year periods. The findings revealed that the accuracies of the Support Vector Machine classification exceeded 0.9, making it suitable for further analysis. Thematic maps generated from the study visualize low green connectivity with poor spatial patterns of green patches in the west of the Urban West Region, primarily due to the higher density of buildup areas. The analysis also indicates the absence of green corridors to enhance connectivity between the patches. Additionally, approximately 0.019% of green area coverage was lost between 2009 and 2018, attributed to shoreline damage along the coastal zone of the eastern side of the Urban West Region. The transition of green spaces, such as trees, shrubs, and grass, into low-density buildup areas and, subsequently, into high-density buildup areas was significant. This transformation poses potential challenges, including increased air pollution and mental health concerns. To address the green challenge issues, the Urban Municipalities of Zanzibar must implement robust strategic plans to preserve and enhance Urban Green Space; such initiatives are essential for promoting sustainable urban development in Zanzibar and mitigating the adverse effects of urbanization on green spaces and overall environmental quality.