Design And Implementation Of Advertisement Data Analysis System Based On Druid

After so many years of development,Internet advertising has achieved very significant results.Internet advertising can help advertisers collect the effectiveness of ad serving,such as potential customer groups and actual user engagement behavior.With the rapid development and rapid advancement of mobile-end apps,traditional PC-based Internet advertising,such as Baidu search advertising and contract advertising of some website media,has gradually been replaced by mobile Internet advertising.Because the Netease news client generates a large amount of advertising log data every day,this brings great challenges to the analysis of advertising data.It is a common problem that the analysis process is time-consuming and incomplete analysis is only based on the existing analysis and processing of data by means of text..Therefore,it is necessary to design an advertisement data analysis system to analyze and process the existing information flow advertisement logs to support the increasing demand of the advertisement business,and provide a reference data basis for the advertisement delivery.This paper analyzes the development of mobile Internet advertising mobile phone application,and combines the background of NetEase news flow advertising business to obtain the functional and non-functional requirements of advertising data analysis system.The functional requirements of the advertising data analysis system mainly include log management,data information processing and statistical analysis of data.The functional requirement analysis mainly uses the use case diagram to model the requirements of the advertising data analysis system.Non-functional requirements include system security and performance requirements.One of the most important performance requirements of advertising data analysis systems is real-time and fast query of data.Druid is a typical representative of big data query engine,which uses high compression and pre-processing of raw data.The way of aggregation can realize the second-level query of billion-level data,which becomes an important choice for industry to analyze and query data for OLAP.Based on the functional and non-functional basis,the architecture design of the advertising data analysis system is derived.The architecture design includes the big data technologies such as Hadoop,Spark and Druid on the technical architecture,and the advertising data analysis is obtained based on the architecture design.The specific implementation process of the system.The architecture design of the advertising data analysis system is mainly from the overall design,the architecture design includes the business level architecture design and the overall architecture design,the system function module design and the system function sub-module perspectives are discussed in detail.The system database design mainly uses Hive storage,giving The meaning of each field in the table and the relationship between the table and the table.In the design and implementation process of the advertisement data analysis system,the advertisement data analysis system is designed and implemented according to the definition of the function module and the business requirements using the class diagram and the sequence diagram.Finally,the advertising data analysis system is tested and verified.In the test part,the hardware environment and software environment required by the advertising data analysis system are first given.The functional test and non-functional test of the system are given in combination with the test environment.The functional test is to verify the actual function module effect through the test case.And combined with the front-end interface gives the page prototype of the actual advertising data analysis system.Non-functional testing is primarily about verifying the ad data analysis system in terms of performance and security.Through actual testing,it is confirmed that the advertising data analysis system basically meets the actual functional and performance requirements.