the school of public policy and management (sppm), tsinghua university released the 2016 big data report on chinese think tanks on january 7, 2017. this report provides a ranking system on the basis of analyzing big data on the activities and influences of think tanks and their social media experts. for the first time, a research institute evaluated and ranked social media big data of chinese think tanks. this report was spearheaded by prof. xufeng zhu, associate dean of the sppm, a leading expert on chinese think tanks.
the 2016 big data report on chinese think tanks mainly measures the influences of think tanks from three aspects: influence of wechat official accounts of think tanks, think tank citation in wechat public opinion, and influence of think tank-verified experts in weibo. the research team selected the most prominent 510 chinese think tank samples from worldwide think tank databases. the team gathered the activity data of these selected think tanks on wechat and weibo by applying data mining techniques on big data. the team then constructed the think tank big-data index, a two-level indicator system, including 3 first-level and 16 second-level indicators, to comprehensively evaluate the influences of think tanks.
the data are collected from over 110 thousand websites, 18 million active wechat official accounts, 150 million active weibo accounts, 6,155 online-forums, and 930 thousand news apps for smart phones. the research team initially captured and gathered data from the abovementioned platforms and then worked on matching accuracy and data selection. the entire data set covers the period from january 1 to november 20, 2016.
the data acquisition steps are as follows:
1. recognize 149 valid think tank wechat official accounts among 18 million accounts from all fields. gather all information on these accounts and over 39,000 released articles that aggregate 78,607,000 "views" and 379,000 "likes" in total.
2. collect 3.84 million articles, which cited the selected 510 chinese think tanks, out of all the articles released by all the 18 million accounts in the year and extract fields from data.
3. use fuzzy identification method to extract 4,516 verified accounts from over 150 million weibo accounts under real-time monitoring. accurately identify 529 experts from 143 think tanks and finish data field extraction of their accounts and articles—151,000 weibo posts, which aggregate 2,663,000 "shares", 1,908,000 "likes", and 1,093,000 comments in total.
according to the overall think tank big data ranking, the top 10 think tanks include china association for science and technology, chongyang institute for financial studies of renmin university of china, party school of the central committee of c.p.c., chinese academy of social science, pangoal institution, center for china and globalization, china finance 40 forum, china business network research institution, liaowang institute, and china center for information industry development.
evaluating think tanks based on big data breaks the technical bottlenecks of objective evaluation and provides a new dimension to recognize the construction and operation of think tanks. this evaluation can also track public opinions in real time to offer reference for governmental decision making.
the influence of big data on think tanks does not equal to the overall policy influence of think tanks. only within a transparent decision-making system can the influence of think tanks be evaluated accurately. therefore, the desire to establish an open and transparent governmental decision-making system fundamentally drives the exploration for a comprehensive evaluation of the influence of think tanks.
download the report:
big data report on chinese think tanks (english).pdf
big data report on chinese think tanks (chinese).pdf
school of public policy and management, tsinghua university
team leader: xufeng zhu, professor
key team members: jing zhao, wanqu han