CAREER: Enabling Reproducibility of Interactive Visual Data Analysis

职业：实现交互式可视化数据分析的可重复性

• 批准号: 1751238
• 负责人: Alexander Lex
• 金额: $ 51.22万
• 依托单位: University of Utah
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2018
• 资助国家: 美国
• 起止时间: 2018-04-01 至 2025-03-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1751238&HistoricalAwards=false
• 关键词: CAREER; Enabling; Reproducibility; Interactive; Visual

Reproducibility and justifiability are widely recognized as critical aspects of data-driven decision making in fields as varied as scientific research, business, healthcare, or intelligence analysis. This project is concerned with enabling reproducibility and justifiability of decisions in the data analysis process, specifically as it relates to visual data analysis. Visualization is an important tool for discovery, yet decisions made by humans based on visualizations of data are difficult to capture and to justify. This project will develop methods to justify, communicate, and audit decisions made based on visual analysis. This, in turn will lead to better outcomes, achieved with less effort and cost. The increasing use of visual analysis tools for decision making will make data analysis accessible to a broad variety of people, as visual analysis tools are generally easier to use than scripting languages and do not require extensive computational and statistical training. This research and its related activities increase accessibility and enhance the data analysis infrastructure for research and education. To achieve these goals, this research will develop a framework for making visual analysis sessions not only reproducible but also reusable. The approach is based on tracking semantically meaningful provenance data during an interactive visual analysis session. Once a discovery is made, analysts can use this history to curate a succinct analysis story, adding justifications and explanations to make their analysis reproducible by others. Using a semi-automatic process, analysts will be able to make their actions data-aware, so that their analysis processes become robust to changes, such as updates in the data. A second contribution of the proposed work is the integration of visual analysis into computational analysis processes. While visualization is commonly used to present computational analysis results, the results of a visual analysis session are rarely used to feed into further computational processes. The techniques developed in this project will allow analysts to feed analysis results (selections, aggregations, filters, etc.) back into a computational environment. This will make it possible to use interactive visualization at any point in the data analysis process while maintaining reproducibility and enabling reuse. The expected results include new methods to capture user intent, create data stories from analysis processes, and to integrate computational and visual data analysis, leveraging the strength of both, human abilities and computational power. The results will be disseminated in publications and in the form of open source software, and accessible via the project website (http://vdl.sci.utah.edu/projects/2018-nsf-reproducibility/).This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

在科学研究、商业、医疗保健或情报分析等领域，可重复性和合理性被广泛认为是数据驱动决策的关键方面。该项目涉及在数据分析过程中实现决策的可重复性和合理性，特别是与可视化数据分析相关的决策。可视化是发现的重要工具，但人类基于数据可视化做出的决策很难捕获和证明其合理性。该项目将开发方法来证明、沟通和审核基于可视化分析的决策。反过来，这将带来更好的结果，并以更少的努力和成本实现。越来越多地使用可视化分析工具进行决策将使数据分析可供更多人使用，因为可视化分析工具通常比脚本语言更易于使用，并且不需要大量的计算和统计培训。这项研究及其相关活动增加了可访问性，并增强了研究和教育的数据分析基础设施。 为了实现这些目标，本研究将开发一个框架，使视觉分析会话不仅可重复，而且可重复使用。该方法基于在交互式视觉分析会话期间跟踪语义上有意义的来源数据。一旦有了发现，分析师就可以利用这段历史来策划一个简洁的分析故事，添加理由和解释，使他们的分析可以被其他人重现。使用半自动流程，分析师将能够使他们的操作具有数据感知能力，从而使他们的分析流程能够对数据更新等变化变得稳健。所提出的工作的第二个贡献是将视觉分析集成到计算分析过程中。虽然可视化通常用于呈现计算分析结果，但可视化分析会话的结果很少用于进一步的计算过程。该项目开发的技术将允许分析师将分析结果（选择、聚合、过滤器等）反馈到计算环境中。这将使在数据分析过程中的任何点使用交互式可视化成为可能，同时保持可重复性并实现重用。预期结果包括捕获用户意图、从分析过程中创建数据故事以及集成计算和视觉数据分析的新方法，充分利用人类能力和计算能力的优势。结果将通过出版物和开源软件的形式传播，并可通过项目网站 (http://vdl.sci.utah.edu/projects/2018-nsf-reproducibility/) 进行访问。该奖项反映了 NSF 的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。

项目成果

Origraph: Interactive Network Wrangling

Origraph：互动网络争论

• DOI: 10.1109/vast47406.2019.8986909
• 发表时间: 2018-12-15
• 期刊: 2019 IEEE Conference on Visual Analytics Science and Technology (VAST)
• 作者: Alex Bigelow;C. Nobre;Miriah D. Meyer;A. Lex
• 通讯作者: A. Lex

Ferret: Reviewing Tabular Datasets for Manipulation

Ferret：检查表格数据集以进行操作

• DOI: 10.1111/cgf.14822
• 发表时间: 2023-06-01
• 期刊: Computer Graphics Forum
• 影响因子: 2.5
• 作者: Devin Lange;Shaurya Sahai;J. M. Phillips;A. Lex
• 通讯作者: A. Lex

reVISit: Looking Under the Hood of Interactive Visualization Studies

reVISit：深入探究交互式可视化研究

• DOI: 10.1145/3411764.3445382
• 发表时间: 2021-05
• 期刊: SIGCHI Conference on Human Factors in Computing Systems (CHI
• 作者: Nobre, Carolina;Wootton, Dylan;Cutler, Zach;Harrison, Lane;Pfister, Hanspeter;Lex, Alexander
• 通讯作者: Lex, Alexander

Taggle: Combining overview and details in tabular data visualizations

Taggle：将表格数据可视化中的概述和详细信息相结合

• DOI: 10.1177/1473871619878085
• 发表时间: 2020-04
• 期刊: Information Visualization
• 影响因子: 2.3
• 作者: Furmanova, Katarina;Gratzl, Samuel;Stitz, Holger;Zichner, Thomas;Jaresova, Miroslava;Lex, Alexander;Streit, Marc
• 通讯作者: Streit, Marc

Predicting intent behind selections in scatterplot visualizations

预测散点图可视化中选择背后的意图

• DOI: 10.1177/14738716211038604
• 发表时间: 2021-10
• 期刊: Information Visualization
• 影响因子: 2.3
• 作者: Gadhave, Kiran;Görtler, Jochen;Cutler, Zach;Nobre, Carolina;Deussen, Oliver;Meyer, Miriah;Phillips, Jeff M.;Lex, Alexander
• 通讯作者: Lex, Alexander