Computational modeling of geospatial data can bring valuable insight into communicable diseases, noncommunicable diseases, and vector-borne diseases. However, each of these disease classes has its own challenges in terms of geospatial data analysis. Complex dynamics of environmental and social risk factors can determine the distribution of some diseases such as COVID-19, obesity and malaria, respectively. Environmental factors such as humidity, precipitation, rainfall, temperature, land use and land coverage, vegetation, and elevation play key roles in predicting prevalence or outbreaks. Measuring these features is time-consuming, costly, and biased since they are subject to human judgment. Remote sensing data, environmental and imagery, could address these barriers by measuring environmental factors consistently and in a timely manner. The novelty of our research study is to develop deep learning models that process variable spatial resolution data for climate, environmental, and anthropogenic features. We will develop multi-scale models to extract deep neural network features that facilitate geospatial analytics at arbitrary spatial extents.