Design And Implementation Of Online Marketing Data Analysis Platform Based On The Materialized Data Cube

In today’s business environment, the competition of Internet group buying market is increasingly fiercer. Group buying market is almost divided into three part:Meituan, Dianping and Nuomi. With high frequent access and operations from mass of users, information systems of enterprises generate a lot of data. How to extract useful information which is helpful to business analysis and decision from these massive data has become an important problems faced by managers.Practical application scenarios of this article are Baidu Nuomi marketing data analysis. Based on the actual marketing big data of Baidu Nuomi, we do multi-dimensional analysis to our marketing effectiveness, which is of great significance to the formulation marketing of strategy and construction of marketing model. According to the effect of different marketing means to purchasing behavior, we select filtering criteria, such as classification of activities, promotional tools, and platform type, to analyze the turnover, new customers’ counts and other indicators of a certain period (experimental or promotional periods). Meanwhile, we build a data warehouse based on users’ order data, browsing data, SMS and app push data. At last, with the Hive platform, we develop a marketing big data OLAP multi-dimensional analysis system, which is convenient for the developers making strategy and operations analyzing activity effect.This thesis uses OLAP technology and the Ares platform, Hive data warehouse, MySQL database and ODP framework to realize data extraction from the layer of source data, design and implementation of the underlying data warehouse server layer and intermediate layer of OLAP server and top layer of front end display.Firstly, the thesis describes the practical application of the system development background, and briefly analyzes data warehousing, online analytical processing (OLAP) and data mining techniques, and sums up the methods and procedures for multi-dimensional data modeling based on OLAP technique.Secondly, consider of actual business requirements of Baidu Nuomi, we establish data-cube dividing dimensions and objects to provide the premise for multi-dimensional data mining of the process. And on the same time, this paper briefly introduces the system architecture of this platform.Finally, this thesis elaborate implementation process of this system, which is thefocus of the article.The innovation of this thesis includes the following aspects: firstly, the thesis based on the actual business processes of Baidu Nuomi, which is of great practical significant to develop the system; secondly, we extract dimensions from business operations to make multi-dimensional data analysis; finally, the system use fan diagram and bar diagram in front-end display so as to show certain indicator trend.