In digital culture, visualizations are a prevalent and ubiquitous form of communication. A veteran journalistic tool, and an increasingly popular one in digital politics, visualizations offer informative value, attract readership, and increase engagement. Visualizations’ multimodality allows them to convey rhetoric through informative, narrative and visual strategies. Arguably, the complex and pliable communicative range of visualizations makes them particularly suited for future-oriented discourse. Indeed, throughout the COVID-19 pandemic, visualizations like the waves of upcoming infections were used to inform the public about possible future realities and relay information about potential next steps and their implications. Despite the rise of visualization-focused scholarly work in the past decade, several analytical lacunas remain, due to visualizations’ multimodal nature and their rich array of actors, contexts and usages in the digital world. Specifically, no scholarly approach examines forward-looking visualizations comprehensively, addressing the ways in which their rhetorical layers coalesce to broker knowledge in multimodal predictive discourse. To fill this gap, our paper proposes a holistic framework for their analysis, addressing knowledge-brokering functions, predictive components, and rhetorical strategies. Thus, we ask, ‘How are predictive visualizations rhetorically constructed to mediate the future?’ and answer through conceptualization complemented by qualitative analysis of predictive pandemic visualizations from journalistic and social media. We combine perspectives from data-journalism studies, projection studies, and visualization scholarship and amalgamate existing tools into an integrative analytical framework that encapsulates each visualization’s rhetorical strategies, its knowledge-brokering functions, predictive structure, and their interrelations. We further refine the tool with empirically-founded specifications and use our empirical applications to offer a new understanding of rhetorical complexity in predictive visualizations. Our analytical framework will serve upcoming studies to examine and define styles of predictive visualization rhetoric across different national contexts, media, and platforms, narrowing scholarly gaps relating to future-oriented visual communication. Read the full OA article here.