Article by Liam Tollinton, Zahra Mahabadi and Yang Peng.
Wednesday 10th October
Bush House, King’s College London
Data visualization expert Mike Brondbjerg came to King’s College to talk to students from the Urban Informatics (Centre for Urban Progress Studies, London) and Geospatial Computation and Spatial Analysis (Department of Geography) courses as part of a series of guest seminars. Mike has a wealth of experience in the fields of data visualization, information design, web applications and generative design.
His talk covered the opposite ends of the data visualization spectrum. One end is about trying to produce insightful and clear data visualization and the other end is about encoding and illustrating raw data in a more creative and abstractive way. He believes that depending on the purpose of a project, the developer should choose an appropriate place along this spectrum for visualizing data in a more effective way. Based on this notion, he put his different projects in a chart to compare them (Figure 1).
Figure 1: Mike’s spectrum of projects
City Hall Projects
He began the talk explaining his current role in the Intelligence and Analysis Unit of the Greater London Authority (GLA), where he works alongside data scientists, economists, statisticians and other analysts. The Unit undertakes a broad spectrum of work, producing the evidence base to inform the Mayor and GLA’s policy and strategy. This includes public opinion research and economic, geographic and demographic analysis and forecasts. The Unit also leads on ‘Smart City’ initiatives which, among many other aspects, aim to harness the power of data to make infrastructure and service delivery adapt better to the needs of citizens.
Mike is building expertise in information design and data visualization, sitting inside the Geographic Information Systems (GIS) team. Part of Mike’s role also involves supporting data visualization capability in the Unit as a whole. He helps to represent findings from analytical work in a more understandable way. Comprehensibility of visualized data is crucial to ensure it has the desired impact.
He introduced two styles of representing data visualization, one of them is more static and the other one is more interactive, allowing people to get involved in exploring the information. In this way, Mike showed some example projects including charts from London Measured (Figure 2), a Population Projection dashboard, and a tool to visualize which areas might benefit most from investment in green infrastructure (due to be launched soon by the GLA) .
Figure 2: London Measured infographic showing key socioeconomic trends in the capital
Mike explained the programming languages, packages used software used in the City Intelligence and Analysis Unit. R and Python, ubiquitous in data science, are used by many of his analyst colleagues for data analysis and simple visualizations. He uses other tools such as Leaflet, VUE/React, Crossfilter, D3, Adobe Illustrator and Processing to produce more complex visualizations tailored to the needs of the project.
Mike has undertaken a wide portfolio of freelance projects as a partner in Kultur Design, for clients across many different sectors. In comparison with projects at City Hall, these have generally been more towards the expressive-abstract end of his spectrum. Mike described the notion of ‘letting data do the drawing’. He said, ‘these projects were not necessarily about insight but more about using data as a material in projects and an input in creative systems’. He talked us through and showed us visuals from some of these exciting projects.
Dow Jones: A dashboard for the Factiva news service of the Dow Jones (the industrial stock market index). He showed us a visualization illustrating the change over time in tech company reference pairs in news articles.
51 Degrees: Visualizing mobile handsets usage around the world on big screen and an app at the 2015 World Mobile Congress, coinciding with the new WebGL release.
Space Oddity: Generative animations of David Bowie’s famous hit Space Oddity with designer Valentina D’Efilippo and researcher Miriam Quick. The final result was animated concentric circles showing harmonies of song, lyric structure and vocal structures by adjusting each instrument to each icon on the circle map.
London Air data: Visualizing changes in Nitrogen dioxide Levels using King’s College air quality data and Open Street Map. The visualization used a moving geometric shape based on sensor locations to show daily fluctuations.
Volanti Imaging: Generative branding work for a drone company. An adaptive logo based on a spiral of points became the centre piece of the company’s brand. Utilising a Weather API, a tool was also developed allowing the atmosphere, pressure and wind speed of London to be visualized within the spiral visualization. In this way, the logo could give a sense of weather conditions in London at that moment.
Tiger Uncaged: A Facebook app allowing customers to generate a unique piece of art work base on their encoding names by mapping icons for each particular letter. Mike noted the results showed interesting cultural differences in naming.
Pulse sensor data visualization: A live visualization (shown in Figure 3) of the pulses taken by people at a series of talks. Blood flow and pressure of audience members were measured as a signal (pulse). Mike observed this kind of data had a sense of dynamics and texture that was important for him to create a drawing and visualizing system.
Figure 3: Pulse Visualization
Papworth Hospital Inpatients Ward: An innovative collaboration with Stephanie Posavec, author of the ‘Dear Data’ book to decorate new hospital in ward in Cambridge. Anonymised patient data on blood flow, spirometer data and others became ‘seeds in generative process’ to create a range of art work.
‘Reasons to’ Logo remix and conference opening titles: Deconstructing the organisation’s logo with music ‘piped through’ to create an impactful visual result. This formed the centrepiece for the conference opening titles.
Generative Art: Various projects including Dead Presidents (Figure 4), Squarepusher Portrait Studies, Generative Still Life and Proximity Studies, which are more abstractive than insightful. In these projects, he tried to display some kinds of artistic emotions and feelings by utilizing raw data.
Figure 4: Dead Presidents
A truly inspiring presentation
At the end of his presentation, Mike fielded a range of questions. These included why he took up a job in City Hall, whether his medical related visualization was suitable for the environment, what advice he would give to people wanting to do his sort of work and how he gets feedback on design work.
Mike quoted a colleague who described data visualization as ‘the art of tidying up’, referring to its normal role of coming at the end of the data analysis workflow. However, his presentation also showed us how data can be a valuable creative spark in the art design process, producing spectator art with a deep meaning.
Mike is at the cutting edge of the data visualization world and we were privileged that he was able to share so much of his interesting work with us. We hope to be able to collaborate with him in the future and thank him for a fascinating talk.
(All images belong to Mike Brondbjerg)