The Music of Data
Presenting data within organizations.
It’s presentation day and you are ready. Your slide deck is polished and all of your key points are backed by customer data in the form of shiny, aesthetically pleasing charts and graphs. You walk confidently into the conference room where the CEO and senior management team are waiting and begin your presentation.
Everything is going smoothly and your presentation is sparking valuable discussion about the sales impact of a newly launched product. But then the sales director asks you to make a quick adjustment to your chart so they can also see the data from an older product for comparison. You have the data, but the chart in your presentation is just a static image taken from your data analytics tool and copied onto the slide…
As your audience watches expectantly, you quickly exit out of your presentation, pull up your data visualization tool, click hurriedly through the different projects there, find the correct visualization, and make the requested adjustment. Not exactly the smooth presentation you were hoping for.
This imaginary scenario illustrates one of the findings of a recent study by Tableau researchers Matthew Brehmer and Robert Kosara: there are tools for giving presentations and there are tools for analyzing and visualizing data, but the two don’t really play well together. The most common way of bridging the gap is by snapping a screenshot of the finished visualization and simply pasting it into the PowerPoint or Google Slides presentation.
But as the above scenario illustrates, this method is limiting because data visualizations are not simply used to pass along information to others. Data visualizations are valuable tools for driving discussions and interactive, collaborative work within organizations. But when those visualizations are reduced to screenshots, the presenter loses that interactivity with the data.
Despite the common use of data visualizations in workplace presentations, relatively little research has been done in this area. Previous research has focused mainly on asynchronous communication channels rather than live presentations with audience interaction. To address this lack of research, Brehmer and Kosara conducted two interview studies with 18 participants from a variety of fields.
In the first study they identified three main scenarios involving the presentation of data which they likened to musical performances:
- Jam sessions;
- Semi-improvisational performances; and
- Recitals.
Jam sessions are casual collaborations between two or more coworkers. Although one person is generally the designated presenter or moderator, the meetings are highly conversational and provide opportunities to explore and iterate with the data. These discussions often involve the use of a data visualization tool so that participants can directly interact with the data.
In comparison, semi-improvisational performances are more rehearsed and structured, although they still provide spaces for improvised interactions. These often take the form of briefings with various departments. Because the audience varies from meeting to meeting, the presenter requires flexibility to respond and tailor the data presentation for each audience.
Lastly, recitals are formal, highly rehearsed presentations given to executive teams or important stakeholders. The presentation is often prepared, revised, and polished by multiple people over significant periods of time, although the presentation itself is typically less collaborative. The data is typically presented in a linear, narrative form.
In the second portion of their study, Brehmer and Kosara prototyped three design concepts and tested them with their interviewees to spark further conversations around specific aspects of data collaboration and presentation.
The first design involved the progressive reveal of data during a presentation. In semi-improvisational performances and especially in formal recitals, the flow of data is typically presented as a coherent narrative. Revealing aspects of the data piece by piece allows presenters to lean into this storytelling approach, heightening suspense and drama to keep their audiences engaged. The prototyped tool gave presenters the flexibility to decide how to reveal the data: they could show each data point one by one, break up line charts to show changes over specific timeframes, or section the data into groups to focus the audience’s attention on different aspects of the data.
In the second prototype, Brehmer and Kosara proposed presenter view controls on a second-screen to give the presenter more flexibility and control over the data during the presentation. By using a second screen not visible to the audience, the presenter could filter the data in response to audience questions or discussions without disrupting the audience’s viewing experience. (Something that would have been helpful in our imaginary, opening scenario.)
The third prototype attempted to make recorded presentations more interactive by enabling presenters to prepare a video recording with data visualizations tied in at different points. The viewer could then interact with those visualizations during playback. Feedback from interviewees revealed that this idea wasn’t broadly applicable to their current methods of presenting, but could be valuable for onboarding and teaching scenarios.
From informal collaboration sessions to polished performances in front of the executive team, data visualizations serve many uses. Brehmer and Kosara’s work shows that there is ample opportunity for improving the tools used to support presentations and conversations around data.
Citation
- Brehmer, M., & Kosara, R. (2021). From Jam Session to Recital: Synchronous Communication and Collaboration Around Data in Organizations. IEEE Transactions on Visualization and Computer Graphics. https://doi.org/10.1109/TVCG.2021.3114760
