Requirements Analysis

Communication University of Zhejiang is a university jointly established by the former State Administration of Press, Publication, Radio, Film and Television and the Zhejiang Provincial People’s Government, and is one of China’s main bases for training radio, film, television, and other media professionals. Over more than 30 years it has supplied large numbers of professionals to central and local media. It currently has more than 13,000 full-time students, with media and arts as the main programs and coordinated development of liberal arts, economics, engineering, and management.

The university has advanced teaching and laboratory equipment and has actively pursued education informatization. In response to the Ministry of Education General Office’s guiding opinions on comprehensively advancing education informatization during the 13th Five-Year Plan, and in accordance with the Zhejiang Education Informatization Construction Project Implementation Measures issued by the Zhejiang Education Department, the university aimed to go further on its existing informatization base and use big-data technology to build a smart-campus big-data analysis and decision system.

The project is to be built in phases, connecting data from all parties at the university into the big-data analysis and decision system, and using the system’s storage and computing capacity for upper-layer business analysis and visualization.

Phase I must ingest data from the academic-affairs, HR, campus-card, library, and wireless-access systems. Raw data are extracted, cleaned, and transformed, then loaded into the big-data analysis and decision system. Vertical and fused analysis is then performed on ingested data—for example, campus-card spending analysis, library-collection analysis, and borrowing-behavior analysis in vertical domains, as well as fused analysis such as comprehensive campus conditions, campus early warning, and faculty/student profiling.

Solution

Based on project needs and Sugon’s big-data strengths, the following comprehensive solution was proposed:

Data from campus systems are extracted, transformed, and loaded into the Sugon big-data analysis platform through Sugon’s big-data integration tools. The platform’s storage, computing, and analysis capabilities support upper-layer business applications. After raw data enter the platform, they are further extracted into thematic databases according to business needs. Analysis results can be visualized directly in the big-data analysis and decision system, or shared with other campus systems through a data sharing and exchange platform.

In upper-layer applications, comprehensive campus-condition analysis presents HR, spending, internet use, and library borrowing from a university-wide view. Campus-card spending analysis covers all spending behavior, from group to individual, and from time-period spending to point-in-time transaction streams. Library data analysis includes static collection analysis and dynamic borrowing-behavior analysis, popular-book recommendation, and correlation of visit counts and duration with academic performance. Faculty and student profiling uses teaching, daily-life, and research data for a 360-degree portrait, identifying group and individual characteristics and enabling personalized management and services. Campus early warning focuses on key individual behaviors, with special attention to students who lose contact, students with internet addiction, and low-spending students from economically disadvantaged backgrounds.

Customer Benefits

Phase I of the smart-campus big-data analysis and decision system enabled Communication University of Zhejiang to ingest selected business-system and machine data. Complex data sorting and preprocessing substantially improved data quality;

A unified big-data storage and compute cluster was established, providing essentially unlimited expandable storage and computing, lifting the university’s IT architecture from a traditional data-warehouse architecture to a distributed big-data architecture and further raising informatization;

Multiple vertical and comprehensive analysis modules were completed, including campus-card spending, library collections and borrowing, faculty/student profiling, and campus early warning. These results let the university grasp campus conditions at any time, instead of collecting data department by department whenever leaders needed a view. Fragmented campus management is made more intelligent as a whole, with more comprehensive information, finer portraits, and smarter decision support.