SmartShots: An Optimization Approach for Generating Videos with Data Visualizations Embedded

Abstract

Videos are well-received methods for storytellers to communicate various narratives. To further engage viewers, we introduce a novel visual medium where data visualizations are embedded into videos to present data insights. However, creating such data-driven videos requires professional video editing skills, data visualization knowledge, and even design talents. To ease the difficulty, we propose an optimization method and develop SmartShots, which facilitates the automatic integration of in-video visualizations. For its development, we first collaborated with experts from different backgrounds, including information visualization, design, and video production. Our discussions led to a design space that summarizes crucial design considerations along three dimensions: visualization, embedded layout, and rhythm. Based on that, we formulated an optimization problem that aims to address two challenges: (1) embedding visualizations while considering both contextual relevance and aesthetic principles and (2) generating videos by assembling multi-media materials. We show how SmartShots solves this optimization problem and demonstrate its usage in three cases. Finally, we report the results of semi-structured interviews with experts and amateur users on the usability of SmartShots.

Designing for Visualization in Motion: Embedding Visualizations in Swimming Videos

“…Much of the literature centers around questions of how to create visualizations in videos to form a coherent narrative and an enjoyable watching experience [41]. Some work, however, has also considered how to embed charts directly in videos [15], [42]. In their design guidelines, Tang et al [15], [42] mention motion factors, but these are primarily related to animating visualizations.…”

“…Some work, however, has also considered how to embed charts directly in videos [15], [42]. In their design guidelines, Tang et al [15], [42] mention motion factors, but these are primarily related to animating visualizations. Yet, several of their design and datarelated considerations are relevant to us even though the authors considered non-sports-related videos.…”