Interactive Progressive Visualization Method For Large-Scale Time Series Data Based On Min-Max Tree Structure

Time-series data is a collection of data points arranged in chronological order,where each data point is associated with a specific timestamp.Interactive visualization of time-series data transforms time-series data into easy-to-understand line graphs and provides multiple flexible interaction methods,allowing users to discover meaningful information from the data in an interactive exploration manner.Therefore,it has a wide range of applications in the scenario of interactive exploration of large-scale time-series data.In recent years,due to the increasing size of time-series data,the visualization of largescale time-series data directly using line graphs in a limited screen will result in a large amount of overlap,which leads to the user’s wrong perception of the data.For this reason,scholars have proposed two different solutions to solve the above problem,one is to enable users to see local data details through interactive techniques,and the other is to use data reduction techniques to reduce the size of the visualized data and avoid the overlap phenomenon.However,when faced with the scenario of large-scale time-series data stored in the cloud database,the latency of querying data in the former will increase significantly,making it difficult to provide users with a smooth interactive experience;the latter is too time-complex or the reduced data cannot generate accurate visualization results,resulting in biased understanding of the data by users.Therefore,the above two solutions are difficult to be applied to large-scale time-series data visualization.In order to solve the above-mentioned problems of large-scale time-series data visualization,this paper innovatively proposes the OM3 time-series data representation based on the minimum-maximum tree structure and the incremental query visualization generation method based on the tree structure.The combination of the two can not only generate accurate visualizations incrementally,but also provide a variety of smooth interaction methods.First,a temporal data encoding algorithm is designed in this paper.The algorithm converts the original data into OM3 representation using maximum-minimum aggregation and difference operations,and then restores the original data by performing inverse operations on the OM3 representation during the interaction process,thus ensuring smooth interaction.Secondly,in order to make the OM3 representation generate accurate visualization results under any size of visualization window and have the ability of incremental visualization,this paper designs a visualization generation method of incremental query based on tree structure.To further improve the efficiency of incremental queries,this paper proposes a node pruning strategy,which improves the visualization generation speed by reducing the size of the query data.Finally,to meet the user’s demand for interactive exploration of temporal data,this paper implements various commonly used interaction methods based on OM3 and proposes various interaction acceleration strategies to improve the smoothness of interaction.To verify the effectiveness of the proposed method,we evaluate the static visualization and interactive visualization with the current state-of-the-art methods for visualizing large-scale temporal data on datasets with different data distributions.The evaluation results show that the method in this paper is not only able to generate accurate visualization results,but also can substantially improve the speed of visualization generation.In addition,we perform a case study of the proposed method based on two real datasets,and the results show that the proposed method is able to generate both a fast overview of the data for the user and a deeper understanding of the local features of the data by providing a variety of interaction methods.