The Research And Implementation Of Time-varying Data Visualization System

With the advent of the era of data explosion,there are a large number of data with complex attributes in various fields,and a large part is time-varying data.Data visualization can graphically display data and help users analyze data efficiently.Many researchers have studied the direction of time-varying data visualization until now.The National Human Genetic Resources Center needs an effective system to visualize and discover the data-varying trends and internal connections,which can help users explore regularity and feature patterns.But there are still some problems.The current time-varying data visualization system(1)is difficult to process and display high-dimensional time attributes;(2)the visualization result is not accurate and clear enough,and current visualization method which is suitable for system applications cannot meet both high accuracy and easy to use requests at the same time;(3)lacks a specific system for time-varying data processing and visualization,and lacks an effective data visualization solution for time-varying data visualization in production environments.These problems need to be solved urgently,so this article has conducted in-depth research on the visualization of time-varying data based on the National Human Genetic Resources Center Data Service Platform.The main contents of this thesis include the following points.(1)For the difficulty of high-dimensional time display,the time series feature representation algorithm is improved,and the fitting error and sequence length are considered comprehensively.In a result,the segmentation efficiency is improved,and the time dimensions are reduced accurately and efficiently.(2)For the problem of unclear visualization result,the K-Means algorithm based on DTW is proposed,which assigns different weights to each time point,makes the intra-class non-similarity as small as possible,and improves the clustering effect.The user can accurately analyze the feature patterns of time-varying data.(3)Design a time-varying data visualization method based on parallel coordinate system,taking the changes of the data itself and time dimension into account,using the timestamp as the coordinate axis.The parallel coordinate system can accurately display the trend of time-varying series data and intra-cluster difference.Based on the above research contents and results,this paper designs and implements a time-varying data visualization system.The system is embedded with relevant algorithms in the background to solve the problem of high dimensional time display difficulty and cluster visualization inaccuracy,which provides users with a complete time-varying data visualization process.The system which is easy to use also integrates multiple visual layout components with good interactivity to help users explore time-varying data implied laws and values accurately and efficiently.