Research And Application Of Multidimensional Data Visual Analysis Based On Log Data

In the era of big data,how to obtain value from massive data are the concern of enterprises,and also the research hotspot of scientific researchers.The traditional data mining and data analysis methods can get a lot of information from the data,but how the information can be understood by users is a difficult problem.Data visualization can present the hidden rules and features in the form of graphics,so that users can quickly and intuitively understand the information in the data,and improve people's cognition and exploration ability of data.In the era of big data,visualization is no longer limited to scientific research and enterprise applications.Interactive visual analysis of data and intelligent computing have become the common basis of major social needs,such as intelligent medical,intelligent transportation,digital industry and so on.In the face of the massive,heterogeneous,multi-dimensional characteristics of big data,the traditional visualization technology has been unable to meet the analysis of these data,so the in-depth study of visualization technology is the need of the times,and the innovation of visualization technology will have a profound impact on the efficient utilization of big data resources.Multidimensional data visualization is a research hotspot under information visualization,and it is a technology to present multidimensional original data or processed data intuitively.How to apply multidimensional data visualization to data intelligence analysis and decision-making is an urgent and meaningful research problem.Based on the online transaction log,this paper studies several problems in the visual analysis of multidimensional data,and puts forward relevant solutions.The main research work of this paper is as follows:1.Visual trend analysis of multidimensional data based on log data.Online transaction log records the relevant information of commodities,users,merchants,trading volume,geographical location,as well as changes over time,which can help analysts understand the sales situation of commodities.The existing visualization methods mainly analyze the purchase behavior from the perspective of users,while analyzing the sales trend of commodities can better help merchants to make business decisions.Based on the transaction log,this paper proposes a framework for trend visualization analysis of multidimensional data and the corresponding data processing algorithm.The concepts of volatility and dynamic performance of data trend are proposed,through which the multidimensional data of time-oriented are displayed in two-dimensional space.The color mapping scheme of data points is designed.The“Feature Ring” is designed to display the data details of individual objects.Based on the above methods,a visual analysis system is designed and implemented.The usability and validity of the visualization methods are verified by using online transaction data.2.Collaborative visual analysis of multidimensional data based on log data.The online transaction data presents significant multidimensional and spatiotemporal attributes.This paper proposes a collaborative visual analysis method for multidimensional spatiotemporal data.Firstly,the multidimensional attribute collaborative visualization view is designed to present the multidimensional properties and their relationships of spatial objects.Secondly,in order to effectively explore the temporal evolution law of multidimensional spatiotemporal data and its implicit feature pattern,a spatiotemporal collaborative visual analysis method is designed.By means of multi-dimensional scaling,the original dataset is mapped to two-dimensional space in a chronological order,which further enable the construction of the sequence parallel coordinates.In order to enhance the visual perception of the parallel coordinates,the coordinate axes are scaled,so the sparse areas are compressed and the dense areas are stretched.When the amount of data is large,there are a lot of curves crossing and overlapping in the parallel coordinates,so users can not accurately identify different objects.In this paper,hierarchical clustering is used to further analyze the display results of parallel coordinates,and the temporal features of different categories can be found more clearly.Through the actual case study of online transaction data,it shows that this method can help users quickly find the feature patterns hidden in multidimensional spatiotemporal data sets.3.Visual analysis of sorting and classification of multidimensional data based on log data.It is a complex task to classify massive multidimensional data,which usually requires iterative experiments on clustering parameters,data features and instances.The number of possible clusters in a data set is sometimes very large,and the exploration of this space is a great challenge.People usually have a more comprehensive understanding of data.For example,they think that data A is better than data B,but they don't know which attributes are important.Therefore,a powerful interactive analysis tool can help people greatly improve the effectiveness of exploratory clustering analysis.This paper provides a visual analysis method to sort and classify multivariate data.Firstly,the user preferences are determined by user's interaction,the weights of each attribute are calculated according to the user's preference model,and then the whole data set is sorted by using the attribute weight set.Finally,the classification is completed according to the sorting results and the user's marking of some data.Through visual display,users can sort and classify data intuitively,and quickly understand the characteristics of data and category characteristics.