Design And Implementation Of General Data Visualization And Analysis Experiment Platform

With the development of computer science and information technology, declared the advent of large data era characterized by Volume,Velocity, Variety and Value.Human society produces a huge amount of data all the time, and these data, in turn, affect the development of human society, and everyone is closely related. So how to analyze the data effectively, understand the data, and then use the data to facilitate people's lives and promote the development of society, is of great significance.Although the traditional data analysis and visualization tools have been fully developed, and many of them are very powerful. However,these traditional data analysis and visualization tools, or only for ordinary users and only provide routine data analysis methods, which can not meet the complex conditions of data analysis or visualization requirements; or only for a specific field, the operation is complex, usually the format of input data is limited, the data analysis algorithm supported is also relatively fixed which leads to a lack of versatility.Based on this, this paper aims to design and implement a general data visualization and analysis experiment platform.The platform is suitable for ordinary users and can meet the normal requirements of data analysis and visualization. It also provides the external interface of data analysis algorithm, which can allow the user to customize the data analysis algorithm according to the specific needs. The platform has nice scalability and versatility.In addition to the design and implement of general data visualization and analysis experiment platform, this paper focuses on the code similarity analysis algorithm based on the syntax tree in data analysis.Based the algorithm, the precision of code similarity analysis experiment is 83.33% and the recall arrives to 78.95% . In this paper, we focus on the visualization of convolution neural network(CNN) in data visualization, and realize the visualization of CNN algorithm based on deconvolution. With this algorithm, the mapping of convolution layer output to the pixel layer is completed successfully in the VGG16 model,and the effect of convolution layer feature extraction is visualized successfully.