In the context of rural revitalization strategy, this paper uses an adaptive density-sensitive similarity measure method to improve based on a spectral clustering algorithm and combines with coordinate visualization technology to cluster rural infrastructure data for visualization. On this basis, a rural infrastructure big data platform is constructed by designing data management, data analysis and data visualization, and the overall performance of the rural infrastructure big data platform is tested in the configured test environment. The results show that when the visualization node is 50000, the running time is only 153.8s, the big data platform can output data stably, the proportion of failure is within 3%, and the corresponding time of each interface on the line is also 1000ms. Based on this research, the infrastructure for a big data platform in rural revitalization strategies can be successfully constructed.