A recent study showed that more than 70% of researchers fail to reproduce their peers's experiments and more than half fail to reproduce their own experiments. Obviously, from a perspective of scientific quality this is a more than unsatisfying numbers. One approach to mitigate this flaw lies in the transparent provision of relevant research data to increase the base of available material to evaluate and possibly reconduct experiments. However, such data needs to be presented and accessed in a findable and purposefully usable way. In this work, we report the development of a programming interface to enhance findability and accessibility of research data (available in DSpace systems) and hence reproducibility of scientific experiments with data. This interface allows researchers to (i) find research data in multiples languages trough automatic translation of metadata; (ii) display a preview of data without download it beforehand; (iii) provide a detailed statistics of the data with interactive graphs for quality assessment; (iv) automatic download of data directly from Python-based experiments. Usability tests revealed that this interface improves the effectiveness, efficiency and satisfaction during the reuse of research data.
翻译:最近的一项研究显示,70%以上的研究人员未能复制同龄人的实验,半数以上的研究人员未能复制自己的实验。显然,从科学质量的角度来看,这不仅仅是一个不满意的数字。减少这一缺陷的一个办法是透明地提供相关的研究数据,以增加可用于评估和可能重新进行实验的材料基础。然而,这些数据需要以可找到和目的明确的方式提供和获取。在这项工作中,我们报告开发了一个编程接口,以提高研究数据的可找到性和可获取性(DSpace系统中可找到的),从而用数据复制科学实验。这种接口使研究人员能够(一) 找到多种语文的研究数据,通过自动翻译元数据;(二) 展示数据预览,而不提前下载;(三) 提供数据的详细统计数据和交互式图表,用于质量评估;(四) 直接从Python实验中自动下载数据。实用性测试表明,这种接口提高了研究数据再利用期间的效率、效率和满意度。