Data Visualization Improvement For Olap Analysis Of Judical Data

In the numerous legal data,using OLAP analysis tools can imitate the user directly multi-angle mode of thinking,for the user to form a multidimensional data model,which support the analysis of the complex operation.In many OLAP analysis tools,Saiku is widely used for its integration of multidimensional analysis techniques and advanced visualization techniques.However,Saiku has the following problems in the process of multi-dimensional analysis and visual presentation of judicial big data.The first one is the process of the mapping of data to graphic elements,Saiku's mapping of geometric tags does not distinguish the categories of data,resulting in inappropriate graphics display.And Saiku's mapping of visual channel in graphic elements is determined by the user,and the system is not filtered and optimized,which results in poor performance in many cases.The second is for the calculation of continuous scale,Saiku using the same method to calculate scale for the different distribution of data,which will lead to unreadable display,affect the user observation;Moreover,for the calculation of the classification scale,because of the arrangement of the dimension of Saiku when dealing with multiple dimensional queries,the problem of mismatch of the comparison object hierarchy can be caused.The last is Saiku lacks interaction in the process of data visualization,so it can't dynamically perceive user requirements and improve iteratively.This thesis proposes an improved method based on graphical grammar to construct the data visualization of the chart.According to the function of the system,the whole OLAP visual display system is divided into two modules: Data processing module and include the design of multidimensional scaling analysis module and data module,among them,the multidimensional data analysis module to complete the task is to use the MySql database to build support star model and coding of the multidimensional analysis model files,convert data in a MySql database into multidimensional data cube to realize the user query analysis function;Data scale design module is the bridge of data to the graphic mapping,before the visual encoding,multidimensional analysis of the result set the module to complete the class work,and according to the different data of different data types is given.Visual coding module is to establish a series of data in the data field to the geometric tag and the mapping of the visual channel,according to the graph grammar bottom-up organized,respectively,to build the data fields to geometric marker mapping graph grammar and data fields to visual channel map graphic grammar.Finally the drawingmodule of the graph organizes the elements in the diagram in a hierarchical way,realizing the reasonable mapping of the data to the graph.The experimental results show that the system the shortcomings of Saiku in data visualization are solved.The system is more flexible in the mapping process of judicial big data.