Combining Lifelines data with creativity to create compelling insights

How design opened the door for scientific data at the Stedelijk Museum Amsterdam

For the past decade, we at CLEVER°FRANKE have challenged ourselves to visualize weather data in a series of posters, which we called the Weather Charts. This series consists of five editions, in which we explored the weather from different angles. Each chart pushes what is possible with open data, data science, visual design, data visualization and printing techniques.

C°F started researching the weather because we think that it poses complex and exciting possibilities to tell stories through the combination of art, design and science. Yet, it is striking that such a high volume of data, which has a significant impact on several aspects of our life, is presented merely through weather forecasts. From the first to the last edition, we focussed on a range of features, from an overview of each day’s weather throughout the year, to social media sentiment analysis in relation to the weather and ranking countries by temperature.

Correlations between weather and everyday life

When we started with the Weather Charts, we realized soon enough that our ambition was much bigger. We wanted to delve deeper into how the weather influences people’s lives; to quantify its impact across various everyday activities and incidents such as sports, crime, and commuting.

To help research this C°F approached Lifelines, a research institute in the Netherlands that studies the health of 167,000 Dutch volunteers in as much detail as possible. With this unique opportunity, we assessed a tremendous amount of data relating to people’s health such as blood pressure and weight, but also their eating and drinking habits, and how they spend their recreational time. Despite the challenges of delving into such a comprehensive data source, ultimately we were able to demonstrate how design and scientific data can be harnessed to communicate complex topics with a broader public.

We started our research by posing several hypotheses and finding data sources that we could align with a weather data set. These data sets were selected based on whether they have the right time frequencies and locations. During this process we needed to clean out the data sets so that we could single out the relevant factors to look into. Afterwards, we harnessed R to create plots and see if the factors that shaped our hypotheses were correlated. Sometimes we couldn’t see a pattern that we expected, which resulted in inverse correlations. For example, we thought that as the weather gets colder and rainy, car accidents would increase. But it turned out to be the opposite, which led us to do further research. We found out that people tend to become extra careful and take more precautions in traffic when the weather conditions are adverse. Ultimately, we’ve curated our findings to showcase both positive and inverse correlations. Among others, these findings include:

• With every 5°C rise in temperature, tea consumption decreases, on average, by 2,1% while coffee consumption drops only 0,4%.
• For every two extra hours of sunshine, the number of days spent gardening per week increases by 13% on average.
• As the weather gets warmer, more ice cream, salads and ready-made meals are consumed.
• The average weight of men and women is greater during the colder months, although men’s weight varies by a larger amount.
• Even blood pressure drops for every 5°C increase in temperature.

While confirming some obvious assumptions on weather and everyday life, our findings also show interesting correlations which may not be ‘groundbreaking’, but pose nuances to our preconceived ideas about the weather’s impact on our life.

Design choices

Every small detail in the Weather Chart is the result of a careful and conscious selection process. We decided to use a square shape as the basis of this chart’s visual language because it enabled the design to adopt a modular system and achieve visual coherency overall.

Through squares, we could ‘chop-up’ maps in different sizes so that people would get a clear idea how many zoom levels we created. For instance, the rougher the shape of the map, the more zoomed-out the view, and the more detailed, the more zoomed-in. These zoom levels also defined the grid for the scatter plots which are used in some sections of the chart.

To support the consistency of the design, we used only three colors: dark blue-silver, fluorescent orange and bronze / gold. The basic information is printed in silver so that the fluorescent orange stands out clearly to show all the data points. The highlights in each section are in bronze / gold squares, functioning as a visual cue to help the reader find an explanation of the data visualizations easily.

Data design challenges

This weather chart was particularly challenging in terms of its aesthetics. Combining a tremendous amount of data in many different shapes and formats in a single piece, and still being able to guide people around the story in a logical way was a difficult goal to achieve. The overall complexity and density of the information can be intimidating, even overwhelming to an uninformed person. The previous Weather Charts possibly worked better aesthetically even though they did not consist of as many different data zoom levels as this one.

Since we wanted to conduct a comparative analysis between the US and the Netherlands, open data sources posed some challenges as well. Both the uncertainty and the occasional lack of open data for certain topics made it particularly difficult to build correlations. For example, when we wanted to see the relation between the weather and the frequency of catching the flu, we could not find this data set in the Netherlands. The scope of our research was sometimes restricted by the availability of open data sets in both countries.

Creatives can uncover insights as long as they’re able to define a concept and know what they are aiming for. Despite these challenges, when looking at the Weather Charts in retrospect we’re satisfied with the valuable insights we’ve gained, highlighting the possibility for future collaborations among designers and scientists. By using our creative power, we presented these insights in a visually compelling way, and received remarkable recognition from the Stedelijk Museum in Amsterdam who included our Weather Charts in their collection earlier this year.

Lifelines’ scientific data has travelled to another realm, where it will inspire a broader public, and hopefully motivate more people to take initiative in finding creative ways to use different information sources and discover new perspectives on data.