The Study On The Platform Of Music Data Analysis Based On Visualization Technology

As the Internet era comes,it promotes the booming development of online music,which also brings the explosive growth trend for the number of digital music and users.The visualization technology is adopted in this thesis to carry out the data analysis according to the massive data resources of NetEase CloudMusic.Besides,the potential rules behind the attention behavior of users is explored,the topic information related to music is excavated,the style types of music and the preferences of users are investigated,so that the data analysis platform of NetEase CloudMusic is established.NetEase released a music Internet product--NetEase Cloud Music on April 23,2013.Its main feature is personalized recommendation service,recommending songs by creators with music aesthetic experience and lovers with independent music taste in the model of playlist;it also includes social function and interactive function.In this era of advocating personalization,NetEase Cloud Music has abandoned the Hot recommendation of traditional music platform,offering a personalized music recommendation service for different people,thus winning massive user resources and music resources.In view of the above situation,this project designs a data analysis platform based on NetEase Cloud Music,mainly including four modules:data collection,data storage,data analysis,and data visualization.The four modules are analyzed by machine learning and deep learning algorithms based on the massive users and music resources of NetEase Cloud Music.The first is user influence analysis,using PageRank algorithm and TriangleCount algorithm;the second is music style classification prediction,mainly using the MLPC multi-layer perceptron classifier algorithm in deep learning;the third is topic mining from comment data,using the community detection BGLL algorithm based on modularity gain.This part contains the main content of this project.A data analysis platform is realized according to the design idea.It collects data with web crawlers and Flume collection technology,stores the collected data on the distributed file system HDFS,then uses Spark distributed computing technology,machine learning and deep learning algorithms to analyze the data,and finally stored the analyzed data results in MySQL and displays the data visually.