Abstract: 【Purpose/significance】A scientifically reasonable evaluation system for scientific research talents in universities can pro⁃ mote scientific research activities and motivate researchers. This article attempts to clarify the current situation of the evaluation of sci⁃ entific research talents in universities in the era of digital intelligence, in order to explore incentive strategies for scientific research tal⁃ ents in universities.【Method/process】This study first uses text mining technology to analyze data on topics such as breaking the "Five Only" evaluation criteria for scientific research talents in universities on the Sina Weibo platform. By constructing a high-frequency semantic network and categorizing the emotions of netizens, the social status and ideological basis of scientific research evaluation in universities are clarified.【Result/conclusion】The data analysis results indicate that universities should integrate multi-source data to strengthen the precise evaluation of scientific research talents, use intelligent technology to assist academic performance evaluation, and use data-driven methods to establish a sound academic incentive mechanism【. Innovation/limitation】Guided by data science, se⁃ lect text data mining and bibliometric methods to obtain relevant data on the evaluation of scientific research talents in universities from online and academic database literature, analyze the evaluation of scientific research talents in universities from different source databases and visually display knowledge graphs, but the sample platform is relatively small.