CAREER: Visually Encoding Personal Data for Vulnerable Populations

“我昨晚睡得怎么样？我现在有多放松明天我会感觉更好还是更糟？“这些问题越来越多地可以通过健身追踪器和其他传感器收集的个人数据来回答。许多支持这种自我认知的工具都依赖于数据的可视化来帮助用户理解它。在设计这些可视化时所做的选择通常集中在如何显示数据，哪些数据被赋予优先级，以及如何比较不同类型的数据。所有这些决定都影响了如何通过个人数据工具创建和支持自我知识。对于具有复杂社会身份的人群（例如患有严重精神疾病的人或边缘化社区的成员），应用于个人数据的视觉惯例通常会对什么是“正常”和什么是“进步”进行有问题的假设。 例如，用于可视化表示数据的常见数据科学技术，如统计平滑，规范基线和连续时间模型，通常与情绪或认知障碍患者的日常生活体验不匹配，这些患者将自我跟踪作为一种治疗形式。该项目将为个人数据可视化开发设计方法和建议，以更好地满足人们身份的独特需求。 这项工作还将产生非歧视性的数据可视化实践、教学材料和专业资源，以更好地反映个人数据可视化设计的方法。这项研究和相关的教育活动将侧重于对特别容易受到歧视性和有偏见的技术影响的人群的个人数据进行视觉编码。经验设计研究将涉及两个弱势群体：1）跨性别和性别歧视的年轻人，他们有患精神疾病的风险，他们的身份往往与个人信息系统中常见的性别假设不匹配; 2）自闭症年轻人，他们的社交活动和视觉化可能与神经正常的人有很大的不同。一系列接地设计调查将为HCI，可视化和数据科学社区提供：1）非专家数据实践如何通过与可视化交互形成的证据; 2）参与式设计活动和协作视觉启发协议; 3）分析个人数据的非歧视性视觉编码的机会，影响和影响;以及4）为弱势群体开发和评估中等保真度的个人信息系统原型。教育活动将对数据科学专业的学生进行以人为本的设计方法培训，为未来的数据科学家提供工具，以减轻他们进入劳动力市场时的歧视性设计做法。教育评估将侧重于这些方法的影响，以帮助学生在处理弱势利益相关者的数据时，反对偏见并做出负责任的决定。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

"How well did I sleep last night? How relaxed am I right now? Will I feel better or worse tomorrow?" These questions are increasingly answerable with personal data collected through fitness trackers and other sensors. Many of the tools that enable this kind of self knowledge rely on visualizations of the data to help users make sense of it. Choices made in the design of these visualizations typically focus on how data are displayed, which data are given priority, and how different types of data will be compared. All of these decisions impact how self-knowledge is created and supported by personal data tools. For populations with complex social identities (such as people experiencing serious mental illness or members of marginalized communities), visual conventions applied to personal data often carry problematic assumptions regarding what is "normal" and what "progress" looks like. For example, common data science techniques for visually representing data such as statistical smoothing, normative baselines, and continuous temporal models often do not match well with the lived day-to-day experiences of people with mood or cognitive disorders who self-track as a form of therapy. This project will develop design methods and recommendations for personal data visualizations that better match unique needs of people's identities. The work will also result in non-discriminatory data visualization practices, teaching materials, and professional resources for more reflective approaches to the design of personal data visualizations. This research and related educational activities will focus on the visual encoding of personal data for populations particularly vulnerable to discriminatory and biased technology. Empirical design research will engage two vulnerable groups: 1) transgender and gender nonconforming young adults, who are at risk for mental illness and whose identities often badly match common gender assumptions in personal informatics systems and 2) autistic young adults, for whom social activities and visualizations may be quite different from neurotypical individuals. A series of grounded design inquiries will provide HCI, visualization, and data science communities with: 1) evidence of how non-expert data practices are formed through interacting with visualizations; 2) participatory design activities and collaborative visual elicitation protocols; 3) analysis of opportunities, impacts, and implications of non-discriminatory visual encodings of personal data; and 4) development and assessment of medium-fidelity prototypes for personal informatics systems for vulnerable populations. Educational activities will train data science students in human-centered design methods to provide future data scientists with tools to mitigate discriminatory design practices as they enter the workforce. Educational assessment will focus on impacts of these methods to prepare students to advocate against biases and make responsible decisions when working with data of vulnerable stakeholders.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.