With the advancement of technology, data visualization has become one of the research hotspots. In big data analysis, data visualization is a very important means, and through visualization software, data related to specific changes can be easily discovered. This article takes the remote dynamic loading algorithm of data as the research object, the alarm system as the research object, and the data dimensionality reduction as the foundation. It studies the visualization environment of Binary Classification data, and studies the visual design, implementation, and visualization process of the remote dynamic loading algorithm of data. The research results indicate that the analysis results of the paper’ system in terms of accuracy, recall, and F1 value are superior to other algorithms (the accuracy of the paper’ system is higher than 80%, with a minimum recall rate of 77% and an F1 value of 0.85).