INTEGRATION OF GIS, REMOTE SENSING AND MACHINE LEARNING FOR THE ASSESSMENT OF ENVIRONMENTAL IMPLICATIONS OF INDISCRIMINATE SOLID WASTE DISPOSAL IN AKINYELE LOCAL GOVERNMENT AREA, OYO STATE, NIGERIA

Indiscriminate solid waste disposal poses severe threats to soil, water, and air quality, yet in many rapidly urbanizing areas of sub-Saharan Africa, the spatial extent of this risk and residents' awareness of it remain poorly quantified. This study assesses the environmental and public health implications of indiscriminate solid waste disposal in Akinyele Local Government Area, Oyo State, Nigeria, by developing an integrated Geographic Information System (GIS), Remote Sensing (RS), and Machine Learning (ML) framework to model household-level risk and its relationship with public awareness.

Sentinel-2 imagery was used within a GIS environment to map waste dump sites and quantify their proximity to roads and natural water bodies, revealing a marked concentration of open dumping along both. Five spectral indices computed from Sentinel-2 reflectance bands, namely NDVI, SAVI and BSI (vegetation health and bare-soil exposure) and NDWI and NDTI (surface water presence and turbidity), were combined with household socio-economic survey data into a multi-layered dataset. Three machine learning algorithms, namely Random Forest, Decision Tree, and Linear Regression, were trained to classify household-level risk and to model the drivers of residents' subjective risk awareness.

The risk-prediction models achieved exceptional performance, with R² scores up to 0.9984, and the risk classification model reached 94% accuracy. Feature importance analysis showed that socio-economic factors, particularly education level (importance score 0.37) and monthly income (0.26), were far stronger predictors of awareness than the objective, model-predicted risk score itself (0.15), revealing a significant disconnect between actual and perceived risk.

These findings identify high-risk, low-awareness households as the most vulnerable group requiring urgent, targeted intervention. The resulting framework offers local authorities a scalable, data-driven tool for designing proactive waste management strategies, including awareness campaigns and infrastructure investment, in support of sustainable development and public health in the Akinyele Local Government Area.