With the improvement of sensor network and human observation capabilities, a large amount of data with spatal and temporal attributes has been generated. Cluster analysis of spatiotemporal data is also a hot topic in recent years.
Therefore, we propose a regionalization-based clustering algorithm considering spatial and temporal contiguity (STR method). In addition, we also integrate it into a toolbox of ArcGIS Pro for users to use.
This figure illustrates the implementation of the spatiotemporal constraint regionalization algorithm, which contains three steps and one iteration rule namely Search Neiborhood, Value Distance, and Neiborhoods Merage.
The following figure is the interface of toolbox.
The Input cubes is input data, 'Attributes' is the attributes users need, 'Number of Spatiotemporal Domain' is the cluster number you want.
The following figure is the regionalization result of synthetic dataset using STR method. K represents the different clustering numbers.
The following figure is the regionalization result of real-world dataset (Chinese air pollutant data in 2019) using STR method.
