With the advent of the era of big data, the generation and accumulation of scientific data has increased exponentially. As the information center for storing information knowledge and providing information services, professional libraries must take the initiative to carry out strategic innovation to meet the demand and create the future.

1. Establish a new digital library integrating data and documents.

The success of the paradigm shift from science to data-intensive science marks the formation of a new conventional science, which will inevitably lead to new breakthroughs and new developments in scientific research concepts and methods, and will generate new demands for professional libraries. Therefore, it is necessary to establish a new digital library integrating data and documents, and form an interoperable framework integrating data and information, so that the whole process of scientific research can be carried out in the electronic environment of the digital library and open to all, so that the materials, ideas, processes and conclusions of scientific research can be spread and enjoyed. Training data management talents in data-intensive scientific research environment "Data is the oil in the information age", and data management talents are scarce talents in data-intensive scientific research environment. Massive data from all over the world are constantly being collected in the United States (or American companies), and this trend shows no signs of changing in the short term. In the future, a country's core competitiveness will largely depend on the speed and ability to transform data into information and knowledge, which actually depends on the technical capabilities of big data. To maintain a leading position in scientific research, national policy makers and researchers must pay close attention to the trend of big data. In its report "Long-term Preservation of Digital Data Collection: Supporting Research and Education in the 2 1 century", the National Science Council of the United States raised the question of how to cultivate and support a new group of scientists called data scientists: "Data scientists include information scientists, computer scientists, database and software engineers or programmers, subject experts, data managers, data indexing experts, librarians, archivists and others who play a key role in the successful management of scientific data resources. At present, the United States needs more than 654.38+0.4 million to 654.38+0.9 million researchers with "in-depth analysis" expertise, while the demand for managers with data knowledge exceeds 654.38+0.5 million. The application of big data is an integrated application with high technical difficulty, such as the need to integrate technical achievements in interdisciplinary fields such as artificial intelligence, business intelligence, mathematical algorithms, natural language understanding and information technology. Data scientist is the most attractive position in the next 10 year, and data management talents such as data librarians and data service librarians will be scarce in the data-intensive scientific research environment.

2. Establish a data-driven E-Science service model.

Under the E-Science environment, whether the library can break through from the traditional information service to the knowledge service will be the key to its survival and development. Therefore, we must pay more attention to scientific data and realize the significance of developing scientific data service and improving the ability of organizing and mining scientific data for scientific research and library competitiveness. Professional libraries should provide researchers with the best information and technical services, meet the needs of long-term storage of massive data, and integrate them into the data life cycle of user workflows. The data-driven E-Science service mode will be a new growth point for the development of modern science libraries.