大数据时代，数据成为与物质资产和人力资源同样重要的基础生产要素。海量数据的空前集聚和计算力的指数级增长释放了深度学习算法的生产力，推动人工智能应用的大发展。大数据技术和大数据产业的蓬勃发展为线上互联网和线下实体经济带来了新的增长点。

The big data play an indispensable role not only in a wide range of science fields and engineering applications, but also in the cognitive mechanisms of the sensation, the quantification, the qualification, the estimation, the measurement, the memory, and the reasoning of human beings. This paper reviews the basic studies of the theoretical foundations of the big data science, as well as a coherent set of general principles and analytic methodologies for the big data systems. The cognitive foundations of big data are explored in order to formally explain the origin and the nature of the big data. A set of mathematical models of the big data are created to rigorously elicit the general essences and patterns of the big data across pervasive domains in science, engineering, and society. A significant finding about the big data science is that the big data systems in nature are a recursively typed hyperstructure (RTHS) rather than pure numbers. The fundamental topological properties of the big data reveal a set of denotational mathematical solutions for dealing with the inherited complexities and unprecedented challenges in big data engineering.

Basic researches of big data science have triggered the emergence of mathematical theories of big data systems. This paper presents a rigorous analytic methodology for big data science and engineering known as Big Data Algebra (BDA). The mathematical models of big data science in BDA are formally elicited from common patterns and essences of a wide variety of big data systems. BDA reveals that any big data system is a Recursively Typed Hyperstructure (RTHS) beyond the traditional domain of pure numbers. It leads to a set of algebraic operators for big data modeling, analysis, and synthesis towards the denotational mathematical structure of BDA. The formal principles and properties of big data and their mathematical manipulations provide a theoretical framework of big data science as the basis for applications in big data engineering.

With the growth of data generated by human activities, the scale, the type and the demands for the data visualization have expanded greatly. In the big data era, the data visualization faces many challenges. In this paper, based on the characteristics and the requirements of the big data, and the current research states of the data visualization, the common data visualization techniques are reviewed. Eight important challenges that the data visualization has to deal with in the big data applications are highlighted. The AutoVis, a data-aware interactive visualization design platform, is specially discussed, as well as its applications.

The rapid development of the big data technology brings about the emergence of various big data products. Today, the ecosystem based on various big data products is very large. Behind the prosperity, the current state of the development of the big data products is difficult to understand for users and practitioners. This paper reviews the core technology of the big data products from two perspectives:the data storage and analysis. Based on the results of authoritative evaluation organizations, the current situation of the big data products in the domestic market is analyzed. Looking forward to the future, China's big data product R&D needs the participations of the open source community, the cultivation of compound talents, the product segmentation and the interdisciplinary collaborative innovation.

Since the launching of the series laws and regulations on network security, the audition of network products and services, data security, there are growing needs for big data service providers for compliance implementation. In order to respond to the data security risks sufficiently and ensure the compliance to network security laws and regulations on big data related business operations, this paper summarizes the key definitions of big data security, the major risks in big data security, and describes the big data security objectives, key security protection technology and mechanisms for big data platforms, as a reference to practitioners in the big data industry.

The Google's GFS and Big Table have broken the limitations of the technology of having to use the relational databases to manage the big data in the past decade. A number of open source big data management systems, such as the Apache Hadoop, carry the technology further by developing more matured technologies and applications. This paper reviews the big data management systems in addressing the usage scenarios of the OLTP and the OLAP based on the Apache software, and the state of art of the data storage engine, the data partition, the data replication, the distributed system protocol, together with a comparison of the pros and the cons of the current distributed file system, the key value store, and the time series database.

How to avoid the privacy and information leakage in the massive data is a widespread concern in the field of the big data and the information security. The data desensitization technology is one of the important means to solve this problem. In recent years, with the rapid development of the artificial intelligence technology driven by the big data and the high-performance computing, a large number of innovative methods were proposed, with many challenges to the existing data desensitization methods. This paper reviews the current situation of the big data and intelligent technology development, and the data desensitization technology, as well as the future development trends of the data desensitization technology.

The urban function mix is a key element in the scheme of improving the efficiency and the vitality. Also, it serves as a significant principle in the modern urban planning. Therefore, the determination of the function mix level of a city and the adoption of effective optimization strategies are essential for the construction of human-oriented cities. A new approach is proposed in this paper for the evaluation of the urban function mix based on the multi-source geospatial big data by combining the spatial layout and the actual condition of the utilization together, to incorporate the characteristics of both supply and demand of urban functions. The method is then applied to explore the spatial characteristics and factors of the urban function mix of the region inside the sixth ring of Beijing. The results suggest that urban function is a common phenomenon in Beijing, and it should be considered as a fundamental principle in the urban planning and management. From the perspective of spatial characteristics, the density of the urban function mix follows a distance decay law, which reflects the significant influence of the location and the polycentric structure. Furthermore, the relationship between the mix ratio and the spatial pattern of urban functions in blocks, dominated by different functions, is discussed. Finally, practical suggestions are made to optimize the urban function mix and to improve the urban fine governance for urban planners and decision makers.

2018年1月1日，中国科学院A类战略性先导科技专项“地球大数据科学工程”（以下简称“专项” ）正式立项，2月12日正式启动。专项启动2年来，在基于地球大数据驱动的科学发现、决策支持、技术创新等方面取得系列重要成果。新年伊始，本刊编辑部对专项负责人、中国科学院院士郭华东进行专访。