Objectives Heat map as a popular and mature visual analysis method at present, shows the distribution trend of geospatial data through the change of color shading. It helps users visualize the spatial distribution and data density of geographic vector point data. The rapid generation technology of geospatial vector data heat maps will significantly empower the improvement of the real-time performance and interactivity of spatial decision support.

Methods Aiming at the problem that the computational efficiency of current heat map generation method decreases greatly with the increase of point data scale, a visualization-driven real-time heat map generation method is proposed, which uses pixels as computing units to directly calculate the heat value to generate the final heat map effects. First, the point data is organized hierarchically based on the tile-pyramid to build a specialized spatial index that can support pixel-based computation. Then, the pixel heat generation algorithm is designed, and the neighborhood pixel superposition strategy is adopted to greatly improve the computational efficiency and maintain the spatial distribution characteristics. Finally, a parallel computing framework is designed to realize interactive heat map visualization.

Results Experimental results show that the proposed method greatly improves the efficiency of heat map visualization, and the time required to generate heat maps for tens of millions of scale geographic point data sets is only 13.5% of the existing method, and the heat map visualization interaction can be quickly completed within 0.75 s, thus supporting the interactive heat map analysis of large-scale geographic vector point data.

Conclusions A large number of experimental results show that the proposed heat map generation method has the real-time performance and interactivity, and compared with the baseline method, the visualization efficiency has been significantly improved.