從美國2009年提出《透明與開放政府備忘錄》（Memorandum on Transparency and Open Government），到我國2012年開始推動政府開放資料，世界各國均積極釋出政府開放資料、鼓勵公私部門應用資料創新公共服務與發揮資料價値。隨著資料科學領域逐漸成熟，「資料經濟」（data economy）、「資料交易」（data transaction）、「資料市集」（data marketplace）、資料信託等概念也應運而生，期待在資料深化其應用方式後，能有更完整的服務或產業生態鏈，將資料的利益發揮到最大化。本研究綜整4個國家（英國、澳洲、日本、新加坡）與歐盟資料交易與市集運作個案，歸納、分析其推動背景、發展趨勢、市集推動政策、資料定價機制，並訪談國內23位產官學專家學者。研究發現：資料交易與市集之經營模式均有顯著的差異，顯見在資料交易與市集推動上，目前並無具備一般性的參照模型，唯資料本身之價値與其所衍伸的前端服務（挖掘資料價値）密不可分，因此平臺可提供資料使用者與加値服務使用者合作之媒合機制變得格外重要；最後，資料定價雖可參考經濟與賽局兩模型或途徑來定價，惟因受資料本身之潛在產値與資料領域之差異，事實上也很難訂定出共通的價格原則。綜言之，資料市集與其平臺營運目前仍需搭配領域與應用特質，採取實驗或演進途徑逐步發展。

When the U. S. government proposed the Memorandum on Transparency and Open Government in 2009, it took policies of open data by storm around the world. The government of Taiwan also launched a policy related to open data in 2012 for the deployment of data in the creation of innovative public governance services. With the maturity of data science, extensive concepts and issues have been raised to develop the operable framework of the data ecosystem, such as “data transactions”, “data economy”, and “data marketplace”. This research collects and integrates the data transaction operations across four countries (the United Kingdom, Australia, Japan, Singapore) and the European Union. We then analyze their promotional background, development trends, data marketplace policies, data pricing mechanisms, and the roles played by governments. With individual and group interviews, the authors also summarize the various opinions of 23 experts assembled from industrial, government and academic domains. The study reveals significant differences in the data marketplace promotional models across the various nations. Naturally, at present, the promotion of data marketplaces lacks generally operable models, while the value of data itself is inseparable from the front-end services. Thus, it is ever more imperative that data platforms facilitate cooperation and coordination in their matching mechanisms among data users, value-added service developers and other stakeholders. As for data pricing, although it can be valued according to either economic or game models, there remains inherent difficulties in establishing common pricing and valuation criteria due to the diversity in the potential output value of data itself, and of the larger data application domains. In sum, the operation of data markets and platform still need to match the characteristics of the domains with the characteristics of their applications. This can be developed incrementally with an experimental and evolutionary approach.