TLS: Towards a Macroscope for Science Policy Decision Making

科学政策制定者和其他决策者需要质量上新颖的工具，以帮助他们识别最重要或最有可能盈利并值得他们充分关注的出版物、专利、技术主张、拨款建议或其他发展。他们还需要一种工具，让他们能够分析和挖掘数TB的相关数据，并以一种易于理解的方式呈现聚合结果，从而使他们能够提高决策质量。然而，我们需要的不仅仅是一个工具。设计真正支持科学决策的工具需要对科学政策制定者的信息需求进行详细的描述和确定优先次序，并建立一个概念框架，将政策制定者的信息需求与现有的数据、分析和指标类型联系起来。一旦这些都到位了，考虑到现有数据集以及理论和算法方法的限制，真正满足当今科学政策制定者需求的工具开发就可以继续进行。这项研究解决了三个挑战。首先，对一组具有代表性的科学政策制定者的信息需求进行了详细分析，包括现有的数据、方法和工具。其次，对与科学决策相关的任务进行了理论概念化，将政策制定者的需求映射到将数据转化为可操作信息的理论上有基础和实用价值的处理管道。然后，这种概念化以类似维基的形式在网上提供，以供社区审查和建立共识。第三，一种原型工具--宏观显微镜--旨在可视化科学技术数据集中的结构、模式、趋势和离群值，这些数据太大太复杂，研究人员无法通过直接观察来理解--就像显微镜和望远镜使人们能够看到太小或太远的东西一样。典型的宏观范围支持一套明确的信息需求，例如，确定新出现的研究前沿，或利用图表和地理空间和科学地图，将研究和探索成果领域的资金与出版物和专利相关联。宏观工具的开发将受益于NSF资助的学术数据库(SDB)和数字基础设施外壳(CIShell)，前者提供2000多万条学术记录，后者支持数据集和算法的轻松即插即用以及独立工具的设计。可用性研究是为了评估和优化宏观范围。通过《地点和空间：测绘科学》展览，将向更广泛的受众介绍宏观及其对先进科学测绘技术的支持。日本的研究人员和科学政策制定者将参与这个项目，产生最佳理论和实践的双向流动。科学研究和科学技术管理需要一个共享的术语和科学如何运作的概念化。它需要可用于运行基准测试和算法比较的共享数据存储库，以彻底验证算法、指标和评估程序的组合。这项研究的理论部分，以及可持续发展计划的延伸和宏观的原型设计，将为科学(政策)界创造基本的共同理解和免费可用的网络基础设施和工具。可以使宏观显微镜成为科学政策制定者选择的工具的相同功能，也可以使其对其他潜在用户具有吸引力。例如，学生或公众将能够看到科学的大规模结构，并找到进一步研究的潜在感兴趣的领域；研究人员将能够监测和获取研究结果、相关的资助机会，或在其研究领域内外找到潜在的合作者。项目经理和审查员将使用它来满足他们的个人需求。

Science policy and other decision makers need qualitatively novel tools to help them identify the publications, patents, technology claims, grant proposals, or other developments that are most important or potentially profitable and deserve their full attention. They also need a tool that lets them analyze and mine terabytes of relevant data and presents aggregated results in an easy to understand way, enabling them to increase the quality of their decisions. However, more than just a tool is needed. The design of tools that truly support science policy making requires a detailed characterization and prioritization of the information needs of science policy makers, along with a conceptual framework that links the information needs of the policy maker to the available data, analysis and indicator types. Once these are in place, tool development that truly addresses the needs of today's science policy makers given the constraints of existing datasets as well as theoretical and algorithmic approaches can proceed. This study addresses three challenges. First, a detailed analysis is conducted of the information needs of a representative set of science policy makers including existing data, approaches and tools. Second, a theoretic conceptualization is developed of tasks relevant to science policymaking that maps the needs of policy makers to theoretically grounded and practically valuable processing pipelines that transform data into actionable information. This conceptualization is then made available online in a wiki-like format for community review and consensus building. Third, a prototypical tool--a macroscope--is designed to visualize structure, patterns, trends, and outliers in science and technology data sets that are too large and complex to be comprehensible to the researchers through direct observation--just like microscopes and telescopes make it possible to see things that are too small or too far away. The prototypical macroscope supports a well defined set of information needs, e.g., identification of emerging research frontiers or correlation of funding with publications and patents in an area of research and exploration of results using graphs and geospatial and science maps. Macroscope tool development will benefit from the NSF funded Scholarly Database (SDB)) that provides access to more than 20 million scholarly records, and the Cyberinfrastructure Shell (CIShell), which supports the easy plug and play of datasets and algorithms and the design of stand-alone tools. Usability studies are conducted to evaluate and optimize the macroscope. The macroscope as well as its support of advanced science mapping techniques will be introduced to a broader audience by means of the Places and Spaces: Mapping Science exhibit. Researchers and science policymakers in Japan will participate in this project, generating a bi-directional flow of best theory and practice.The scientific study and management of science and technology requires a shared terminology and conceptualization of how science works. It requires shared data repositories that can be used to run benchmarks and comparisons of algorithms, to thoroughly validate the combination of algorithms, indicators, and evaluation procedures. The theoretical component of this study, as well as the extension of the SDB and the prototypical design of a macroscope, will create a basic shared understanding and a freely available cyberinfrastructure and tool for the science of science (policy) community. The same capabilities that could make the macroscope a tool of choice for science policy makers could also make it attractive to other potential users. For instance, students or members of the public will be able to see the large scale structures of science, and find potential areas of interest for further study; researchers will be able to monitor and access research results, relevant funding opportunities, or find potential collaborators within and without their fields of inquiry. Program managers and reviewers will use it to meet their individual needs.