Data-driven Scientific Research Correlation And Impact Analysis

Science of science(SoS)is the science that study science,aiming at understanding,quantifying and predicting scientific research and its outcomes and impact.SoS reveals rules underlying the development of science and scientific activities,then,applies these rules back to promote the science,assisting scientific development strategies,planning and policy,making science research more efficient.The era of big data provides a wealth of data source and data analysis tools for SoS.Carrying out data-driven scientific research relevance and impact analysis has important theoretical and practical value for the field of SoS.This paper takes the data analysis of NSFC as an example.Firstly,combines Selenium and Tesserocr to compile the web crawler program to obtain the funding and paper data successfully,writes regular expression for extracting the funding imformation in scientific papers to connect funding and paper data;Secondly,analyzes the characteristics of funding paper data to seek for an efficient data management method for better storing not only funding and paper data but also the relationship between them,and finally adopts the graph database Neo4 j,a relationship-centred databased for data management.Thirdly,uses scientometrics,socioeconomics,and correlation analysis to analyze data: taking Gini coefficency to quantitatively assess the imbalance distribution of research funding among scientific research institutions,indicating a big gap in the distribution of funding among institutions;studying the correlation between the funding input and the output of scientific fund from the level of funding project and institutions input,showing that there is no significant correlation between the number of project achievements and the project amount while there is a strong correlation between institutional research funding amount and number of achievements.Finally,constructs scientific collaboration networks in institutional and national(or regional)level and studies the evolution process and cooperation mode of the two kind of scientific collaboration network,showing that the funding input has different promotion effects on the scale and density of these two kinds of networks and the funding level of the organization has a significant impact on the “status” of the organization in the network.The innovative works of this research are the scientific research data management method including scientific research data acquisition and graph database management method isproposed,the imbalance distribution of scientific research funding and correlation analysis between funding input and output in project level and institution level are explored,and the funding cooperation relationship and the impact of funding to nodes importance based on the complex network model is analysed.All in all,this study provides technical support for the acquisition and management of scientific research data and provides scientific funding-paper data analysis method forSoS.Inaddition,the result provides decision support for the science funding management department and promotes the development of SoS.