This feeling was familiar. I felt it when I watched the title credits of Gaspar Noe’s Enter the Void (may contain flashing images). The film follows a man’s drug-induced hallucinations whilst wandering around Tokyo. Neon words flicker on screen, as if lit by strobe lights. There’s a low humming sound in the background. Then a moment of silence.

Suddenly, the silence becomes an unrelenting, pounding beat. Incongruous typography is thrown at the viewer, in time to an increasingly frenetic soundtrack. The whole thing leaves you overwhelmed, lost amongst incomprehensible text and neon. There's a disconcerting feeling of being in a different mental state - a taste of hallucination. In Enter the Void, the opening title summarises the main themes and narrative of the film in just two minutes. In unfold, something similar is happening, but using astrophysics data.

Data as we usually think of them - numbers ferreted away in spreadsheets - are not usually great at making us feel things. Apart, perhaps, from frustration, anger and confusion when trying to figure out what they mean ( just me?). In in their raw form, datasets are often difficult to interpret, analyse or tell stories with. That’s where representing data in interesting ways comes in.

Narrative, emotions and action

Portraying data in creative but accurate ways allows us to have emotional reactions to research findings. It allows a narrative to emerge.

One interesting example is artist Kasia Molga’s Human Sensor Project, which is showing in Manchester at the end of July. This project showcases clothing items which change colour in response to air pollution measurements. It also lights up in response to the wearer’s breathing. The Human Sensor makes pollution and its geographic variation visible. It reminds us of the relationship between pollution and one of the most basic, universal experiences - that of breath. Air pollution can affect our breath but we also inhale pollution without noticing, breath by breath.

An important aspect of emotional responses to data is the inspiration to make change. Many pieces of art are inherently political. They include explicit calls to action. They force the viewer to see or feel uncomfortable things, in the hope that they will change behaviour or thought processes. From Banksy’s depictions of Cosette from Les Miserables as a Calais migrant to Pussy Riot’s high profile performance in a Moscow Cathedral and subsequent criminal trial, artists are not shy of openly wanting to change hearts and minds.

Often data relevant to social justice, equity and wellbeing are available, yet they are not effectively translated into action. This visualisation of gun deaths, described in an excellent blog post from the Centre for Artistic Activism, is a great example of the power of data to inspire emotive responses and (hopefully) action. Much social justice campaigning focuses on making the experiences of individuals heard. Data brings together the experiences of many individuals in a systematic way, creating a clarity that can be useful in inspiring change.

Scientists often shy away from connecting their work to political causes or personal beliefs. Perhaps this is due to a desire for science and the scientific method to be unbiased, neutral. Yet the decision to explore certain questions can and does have political and social relevance. This is not always a bad thing. We cannot study everything, and so we must prioritise. This can be driven by our own motives and interests or by those of funders – whether governmental, not-for-profit or commercial.

Drunks and lampposts: a warning

“Statistics are used much like a drunk uses a lamppost: for support, not illumination.” - Vin Scully, or Andrew Lang, depending on who you believe.

As humans, we tend to rally around stories, and find ways to make statistics match these stories. However when data are the stories, is this different? How do we combine a desire to create a narrative, inspire emotional responses and even action without creating bias? Without misrepresenting? How can aesthetics and accuracy be combined?

For data visualisation to be useful (in the broadest sense of the word), effective collaboration between artists and scientists is needed. This requires humility, and compromise. Sometimes the compromise will need to sacrifice aesthetics, sometimes accuracy. We also need to recognise that there are many people who are talented at and enjoy being both artist and scientist. Generally the best art and the best science involves stories, just told in different ways, using different tools. We can start by trying to tell the stories important to us, but being open minded about how they end.