Data visualizations are used to communicate messages to diverse audiences. It is unclear whether interpretations of these visualizations match the messages their creators aim to convey. In a mixed-methods study, we investigate how data in the popular science magazine Scientific American are visually communicated and understood. We first analyze visualizations about climate change and pandemics published in the magazine over a fifty-year period. Acting as chart readers, we then interpret visualizations with and without textual elements, identifying takeaway messages and creating field notes. Finally, we compare a sample of our interpreted messages to the intended messages of chart producers, drawing on interviews conducted with magazine staff. These data allow us to explore understanding visualizations through three perspectives: that of the charts, visualization readers, and visualization producers. Building on our findings from a thematic analysis, we present in-depth insights into data visualization sensemaking, particularly regarding the role of messages and textual elements; we propose a message typology, and we consider more broadly how messages can be conceptualized and understood.

Top