DashGuide: Authoring Interactive Dashboard Tours for Guiding Dashboard Users

TL;DR

DashGuide tackles the challenging task of authoring guided dashboard content by introducing an interaction-first workflow that records author actions and uses a large language model to generate contextual tour content. The system extracts real dashboard metadata via Tableau Public API, captures targeted interactions, and renders tours as playable, in-dashboard overlays with editable steps and playback. Formative insights from 9 practitioners informed design goals, and a summative study with 12 dashboard creators demonstrated efficiency gains and high-quality content generation, while highlighting challenges around data updates, context control, and multi-tour delivery. Overall, DashGuide provides a practical, author-centric pathway to scalable, expressive dashboard guidance, with clear directions for extending to reuse, personalization, and end-user evaluation.

Abstract

Dashboard guidance helps dashboard users better navigate interactive features, understand the underlying data, and assess insights they can potentially extract from dashboards. However, authoring dashboard guidance is a time consuming task, and embedding guidance into dashboards for effective delivery is difficult to realize. In this work, we contribute DashGuide, a framework and system to support the creation of interactive dashboard guidance with minimal authoring input. Given a dashboard and a communication goal, DashGuide captures a sequence of author-performed interactions to generate guidance materials delivered as playable step-by-step overlays, a.k.a., dashboard tours. Authors can further edit and refine individual tour steps while receiving generative assistance. We also contribute findings from a formative assessment with 9 dashboard creators, which helped inform the design of DashGuide; and findings from an evaluation of DashGuide with 12 dashboard creators, suggesting it provides an improved authoring experience that balances efficiency, expressiveness, and creative freedom.

Figures (3)

• Figure 1: Visual Interface of DashGuide. This example uses the same dashboard as the teaser image but the communication goal of the dashboard tour is data facts instead of dashboard semantics and encoding. a) Buttons for recording interactions and reloading a dashboard. b) On clicking the Record button (), a modal opens where authors can select a communication goal, provide instructions to the LLM (optional), and a title (optional). c) After choosing the communication goal, authors can interact with the dashboard for creating the tour (shown with the hand cursors). d) Based on the interactions, a dashboard tour is created using an LLM. Each step in the tour contains three icon buttons for playing back the original interaction, editing the step, and deleting the step. e) Playback of a step in the dashboard. The title and description appear as an overlay in the dashboard.
• Figure 2: Interfaces for editing a tour. a) The top row in the interface shows the title of the tour as well as four editing features for the tour. The play button () hides all other components in the tool, except the dashboard, and then plays all steps one by one. The edit button () opens the setting modal from \ref{['fig:interface']}b. The other two buttons allow saving () and deleting () the tour. Each step of the tour contains three icon buttons. b) The first button () plays back the interaction associated with the step. c) The second button ([regular]) opens a modal for editing the setting for an individual step. This interface allows users to change the guidance intent for a step, provide further instructions for regenerating the step, and manually edit the title and description. d) Users can insert a new step without any interaction by clicking the plus icon () and then outlining the instruction (e.g., "add a transition step here") for the step. Authors can also click on the record buttons () in between the steps to add new steps with interaction.
• Figure 3: Editing a Dashboard Tour in DashGuide. The dashboard in this example analyzes flavors (e.g., strawberry, peach, etc.) in fruit snacks. Consider that the author created a dashboard tour with the goal of communicating data facts. a) A step of the tour features a jitter plot. Thinking the users may not know about jitter plots, the author regenerates the step with another communication goal, dashboard semantics. b) The step is updated with an explanation on jitter plots. c) The author manually tweaks the description for a clearer explanation.