Visualising climate change, measuring racial segregation, and a defence of weird charts: Highlights from OpenVis Conf

OpenVis Conf brought together people from around the world to learn about data visualisation, design, and analysis. Here are the highlights of the event.

Organised by emlyon business school‘s Data R&D Institute, the two-day event took place in Paris on 14–15 May (followed by a day of workshops). It brought together data visualisation enthusiasts from a wide range of industries to learn from visualisation experts working in the academia, tech industry, and journalism.

Read on to find out what ‘xenographphobia’ are, why disagreements can inspire better visualisations, and how visualisation can be a tool for liberation.

Xenographics: Why we should all be William Playfairs – Maarten Lambrechts

As a ‘hobby research project’, Maarten Lambrechts recently started collecting examples of strange charts, or as he calls them, xenographics (xeno being Greek for foreign or strange). His site features a gallery of examples with descriptions and links to the sources. (You can also submit new charts.)

How does one go about creating original charts? Lambrechts gave a few suggestions, presented in the order of complexity:

  • Designing graphs for mobile screens sometimes results in original charts. Try, for example, flipping the axes of your chart, or rotating it (check how the Parliament composition chart turns in this Financial Times story when you make the window smaller).
  • Include more data in a smaller space than you usually would, for example by creating overlapping bars in a bar chart.
  • Combine different types of visualisations – see for example a tree map fused with a bar chart, a combined line and bar chart, or even the ‘drunken speedometer’.
  • Add new dimensions to existing charts. Hans Rosling basically reinvented the scatterplot by adding two new dimensions: the size of the bubble (value) and animation (time).
  • Finally, for very complex charts, use new ways of visualising spatial movement: for example the OD map preserves the spatial layout of origin and destination locations to construct a two-level spatial treemap. Or have a look at the temporal cartogram, which is a spatial representation of travel time, not the actual distance.

Finally, Lambrechts discussed a related condition that a lot of people suffer from: xenographphobia, or fear of weird charts, ‘Editors in newsrooms are very xenographobic, thinking that their readers won’t be able to understand more complicated visualisations’, he said. ‘I think we need to break away from this. As long as we explain how to read the charts, people will get it’.

Disagreements – Amanda Cox and Kevin Quealy, The New York Times

Amanda Cox and Kevin Quealy’s presentation showed how data visualisations can provoke passionate debate between visualisation experts. Touching on some points of disagreements they’ve had during their time working together at the New York Times, the discussion opened a fascinating view on how the data design team approaches visualisations at the newspaper.

The first disagreement was about simple charts. Quealy said that simple charts are often the most effective way of transmitting information, which should be the priority in all reporting. When examining the opioid crisis in the US, for example, the project collected data from around the country but ultimately visualised it using very simple charts. ‘With all the work that goes into this kind of research, it’s a horrible idea to waste it on some experimental chart’, Quealy said.

To argue against this, Cox brought up the Times’s project that looked into income mobility based on race and gender. This project used a visualisation type that was not intuitive to read, but still very effective.

The journalists also discussed the Time’s recent interactive about the US Tax Bill: Quealy said that most people were mainly interested in the number on the graph that told them if their taxes would go up and down, while the rest of the visualisation was mostly superfluous. ‘But this relates to the idea of signalling: charts that look serious help to signal to the reader that this is a really important topic they should be interested in,” he said.

Do you need to be obsessed with an issue in order to produce your best work? No, argued Cox, ‘Covering healthcare is not exactly why Kevin wakes up in the morning. But turns out he can still make some really great charts about the topic’.

On the other hand, Quealy brought up his interest in Donald Trump’s insults, which was behind one of the most memorable recent visualisations from the Times. ‘I was so obsessed that I would start collecting them. We kept it going, and eventually run it in print – and gave it a full two-page spread’.

Visualising climate change – Nadja Popovich, The New York Times

While being a huge story in journalism, climate change can be a remarkably difficult topic to visualise. Nadja Popovich talked about how the New York Times has handled climate-related visualisations.

One of the most commonly seen visualisations concerns data on how average temperatures change over time. Charts also very often visualise the decline of arctic sea ice.

But the problem with both of these visualisation types is that the topic may end up feeling distant to the reader. To bring the matter closer to home, the New York Times looked at how glaciers in the Glacier National Park in Montana have melted over the past 50 years. ‘This is a place that many Americans had visited, so it was a particularly hard-hitting thing for a lot of people’, Popovich said.

The New York Times has also visualised how extreme weather events such as extraordinarily hot summers are becoming more common, but not all readers were convinced of the coverage. ‘Some of the reactions we got were, ‘Well it’s not getting hotter here’, or ‘I’m cold’. We’re often talking about averages over a 30-year period, and people find that a little bit hard to relate to’. One solution to this is to present the data on maps, which gives readers information about their specific locations.

Finally, Popovich pointed out that it’s not just climate data that can be visualised –the policy behind climate change can also provide effective visualisations. For example, the Times has looked at potential future scenarios, which depend on how effectively societies work against climate change, and visualised different estimated trajectories.

How data, and the visualisation of it, helps us understand ‘US’ – Aaron Williams, The Washington Post

Aaron Williams is behind one of the most praised data visualisation projects in journalism lately, an investigation into race and segregation in the US. What inspired him to look into the topic were charts by W. E. B. Du Bois dating back 100 years, that visualised race and distribution of the black population in the US around the year 1900, ‘It was striking how similarly concentrated the black populations are still today, so I wanted to dive into this idea’.

Williams wanted to focus on segregation specifically, rather than looking at race more generally, like some other journalistic projects have done. This lead him to discover and use a statistical measure called ‘multigroup entropy index’, which shows how uniform the ethnic distribution is for a given area.

The index allowed him to create maps of individual cities, depicting how racially segregated Washington DC, Houston and Chicago and other cities are. ‘But then my bosses were like, just do it for the whole nation’, said Williams.

Williams also spoke to people who research segregation, and they confirmed what his visualisations were showing: despite becoming more diverse, the US is still a highly segregated country.

‘You often hear people say that Chicago is a diverse city. But it’s not’, Williams said. ‘It’s diverse in that you have a lot of different people there, but the maps show that they live in these different concentrations’.

Finally, going back to Du Bois, Williams rephrased his quote about how data visualisation can be used to show societal progress. ‘I would interpret it as: ‘Visualisation can facilitate liberation’. It has the potential to move people, or present really powerful ideas, in ways that pure text or photographs can’t’.

‘So if you are in the business of trying to liberate people, data visualisation is a great way to do it’.