Abstract: [Purpose/Significance] Research on the background, concept, representation and application of data diversity can provide support for the effective data management and governance. [Design/Methodology] The representations of data diversity is summarized mainly by the observation method and inductive method, the concrete data diversity contained in specific applications is illustrated by phenomenological interpretation method. [Findings/Conclusion] It is found that data diversity is the richness & difference of information carried by data and the complexity and variability of types, dimensions, formats, versions, and forms of data. Data diversity has three forms of representations each inside and outside. Data diversity requires subject diversification, collaborative governance, and common rules in data management. Data diversity is already reflected in open science and industries and libraries, and librarians should be aware of this. [Originality/Value] This paper proposes the concept of data diversity and briefly discusses the requirements of data diversity for data management.

Key words: Data diversity, Open Data, Research data, Open Science, Big Data, Data management