报告题目：Unifying Conceptual, Logical and Graph Data Modelling Using Principled Entity/Relationship Graphs

报告时间：2025年12月22日上午10:30

报告地点：85porn B405会议室

报告人：Sebastian Link

报告人单位：新西兰奥克兰大学

报告人简介：Sebastian Link is a Professor of Computer Science at the University of Auckland, where he is also the Director of Data Science in the birthplace of R, the world’s most used language for statistical analysis. As the Associate Dean International for Science, Sebastian has helped establish many transnational education programmes with leading universities in China and elsewhere. His contributions to data management earned Sebastian the 2013 Chris Wallace Award, the most prestigious research award for mid-career Computer Scientists in Australasia. He was also awarded a Doctor of Science by the University of Auckland in 2015. Sebastian’s research spans data dependency theory, data profiling for which he co-founded the company DataViadotto, conceptual and logical database design for a diverse range of data models. He has published in all top database conferences and journals, is currently on the editorial board of the VLDB Journal and Information Systems, and a regular member of programme committees such as SIGMOD, VLDB, and ICDE. Sebastian has led many prestigious research grants in New Zealand to successful outcomes.

报告摘要： Entity/Relationship modeling is a great methodology with a fifty year track record of designing high-quality database schemata. Typically, this happens conceptually before translating the model to the logical level, such as relational databases. The talk will promote the recently introduced concept of Entity/Relationship Graphs, which establishes the first graph semantics for Entity/Relationship models. It will be demonstrated that E/R diagrams constitute a subclass of property graph schemata whose instances are E/R graphs that are designed well, referring to the absence of data redundancy and processing difficulty on target workloads. Furthermore, E/R graphs offer new opportunities in the management of entity and referential integrity. As a proof of concept, we showcase how the famous relational TPC-H benchmark performs when its schema, instances and workloads are translated into Neo4j.

邀请人：魏子恒