Enhancing Web-based Analytics Applications through Provenance
Enhancing Web-based Analytics Applications through Provenance
复制标题
通过 Provenance 增强基于 Web 的分析应用程序
DOI:
10.1109/tvcg.2018.2865039
复制
发表时间:
2019
影响因子:
5.2
通讯作者:
D. Koop
中科院分区:
文献类型:
--
作者:
Akhilesh Camisetty;Chaitanya Chandurkar;Maoyuan Sun;D. Koop
Visual analytics systems continue to integrate new technologies and leverage modern environments for exploration and collaboration, making tools and techniques available to a wide audience through web browsers. Many of these systems have been developed with rich interactions, offering users the opportunity to examine details and explore hypotheses that have not been directly encoded by a designer. Understanding is enhanced when users can replay and revisit the steps in the sensemaking process, and in collaborative settings, it is especially important to be able to review not only the current state but also what decisions were made along the way. Unfortunately, many web-based systems lack the ability to capture such reasoning, and the path to a result is transient, forgotten when a user moves to a new view. This paper explores the requirements to augment existing client-side web applications with support for capturing, reviewing, sharing, and reusing steps in the reasoning process. Furthermore, it considers situations where decisions are made with streaming data, and the insights gained from revisiting those choices when more data is available. It presents a proof of concept, the Shareable Interactive Manipulation Provenance framework (SIMProv.js), that addresses these requirements in a modern, client-side JavaScript library, and describes how it can be integrated with existing frameworks.