Spatiotemporal visual analysis research is rapidly emerging and evolving, as scientists engaging in cross-domain research efforts, introducing novel systems to perform data-driven spatiotemporal analysis in several domains such as astronomy, climatology, environmental science, urban planning, etc. However, due to the increasing volume and complex nature of the spatiotemporal data, scientists encounter challenges to hypothesize, investigate, and compare the spatiotemporal variables. Often the spatiotemporal data are stored in disparate and fragmented forms in distributed data sites. Therefore, important insights may not reside in a single spatiotemporal dataset but rather distributed in multiple remote data sites. Such a scenario introduces a massive overhead in terms of acquiring the data, preparing the computational environment, and conducting exploratory spatiotemporal analysis. Moreover, our literature review and surveying the domain researchers suggest that exploratory visual analysis of spatiotemporal data still significantly relies on static visualizations to present specific data stories. An interactive visual analytics system obtain the potential to benefit in this context, providing a dynamic platform for scientists in performing distributed spatiotemporal data exploration.In this dissertation, we address these challenges to outline the design requirements for a cloud-based visual analytics approach that serves distributed spatiotemporal data exploration. The interactive exploratory spatiotemporal visual analytics approach consists of three major components. First, we present a distributed data mining architecture in a visual analytics framework that supports unified spatiotemporal data access, transformation, and analysis. Next, we present an interactive contour-based geospatial visualization that supports exploratory and comparative geo-spatiotemporal visual analysis. Finally, we present a pipeline for a visual analytics interface that sources distributed spatiotemporal data using the data mining architecture. To support the scientists in exploratory analysis, the pipeline provides interactive contour visualization in coordinated multi-views. To demonstrate the scalability and scientific value of the analytical workflows, we conducted qualitative and quantitative user studies. Results from the user intervention study and domain experts' feedback suggest that the proposed interactive visual analytics approach significantly improves users' performance in performing exploratory analysis over the distributed spatiotemporal data.