User Perceptions of Actionability in Data Dashboards

Abstract

This article reports on a multiphase study designed to understand how nonexpert users interact with COVID-19 data dashboards, particularly in terms of the dashboards’ actionability, or ability to support decision making. Analysis of the videos and transcriptions of user interviews shows the variable relevance of proposed criteria for dashboard actionability and suggests additional criteria for users’ emotional responses to data and for the presentation of data at degrees of personal and local granularity. These findings advance an understanding of how nonexpert audiences interact with and derive value from complex visualized data.

National Public Health Dashboards: Protocol for a Scoping Review (Preprint)

“…Ivanković et al [18], for example, defined data dashboard actionability according to 7 features: (1) knowing and clearly stating the desired consumers of the information, (2) selecting and presenting appropriate indicators, (3) clearly stating the sources of data and methods used to generate indicators, (4) demonstrating variation over time and linking changes to public health interventions, (5) providing as high a spatial resolution as possible to enable consumers to evaluate local risk, (6) disaggregating data to population subgroups to further enable evaluation of risk, and (7) providing narrative information to enhance interpretation of the data by the consumer. This type of user-centered conception understands actionability as a function of both usability and degree of match between data and users' information needs is intuitive but may not be adequate or sufficient to assess the actionability of dashboards intended for a general audience [19]. Other scholars in this space considered a design-centered conception of actionability [20].…”

“…In their view, to be actionable, dashboards must prompt or trigger users to act on data by being integrated, via behavioral design, into users' data use practices or routines such as assessing performance on tasks or progress on goals. Finally, some advocate for a decision-centered conception of actionability, whereby data dashboards are considered actionable to the extent they provide data, analyses, and forecasts (eg, predictive analytics) that allow decision makers to make an informed choice among alternatives [19,21,22]. All 3 conceptions appear to be relevant to the definition and operationalization of actionability as a key design feature of public health data dashboard and the scoping review will be instrumental both in terms of more fully explicating actionability based on the integration of existing conceptions, as well as identifying additional potential dimensions that may be used to this end.…”

Public Health Dashboards in Overdose Prevention: The Rhode Island Approach to Public Health Data Literacy, Partnerships, and Action

“…However, such data dashboards were in their infancy, and the governmental and research organizations creating them had urgent and multidimensional needs. Dashboards were expected to convey rapidly changing disease patterns, monitor emerging issues in local communities, and communicate about the risk of individual exposure [ 4 , 5 ]. Data dashboards are now a common platform for communicating policy changes and displaying trends for decision makers, researchers, and the news media [ 6 ].…”

“…Public health data dashboards created to monitor the overdose crisis and other ongoing public health issues across the United States are no different. Given their growing presence and appeal, the landscape of public health data dashboards leaves room for focused development and improvement in engaging users through accessible data storytelling, sustainable community partnerships, and data-driven decision-making [4,5,[7][8][9][10]. The aim of this viewpoint paper is to describe the Rhode Island approach for data dashboard development, an instructional approach to overdose dashboard development driven by our experiences to date.…”