I made a book. It documents my major self-initiated project this year: a process-based investigation into generative design and the weather. It’s a record of the sky, of the colours that change so slowly it’s almost impossible to see. A bit like a visual almanac.
I linked 64 public-access web cameras across Europe, recording the colour of the sky, at each point, at regular intervals. Together, the cameras paint the weather, once every hour. The book collects a week of paintings.
What began as a project investigating data visualisation, turned into a project exploring painting. Hidden amongst the changing hues, are glitches, imperfections and accidents caused by the lenses themselves. Eyes on the sky, painting by numbers.
This project was inspired by a simple observation: weather conditions can be revealed by the quality and colour of light through a window. A soft, blue light falls in overcast conditions in the winter, when the Sun is weak and far away. On a warm summer day, light is bright and tinged with red.
With regards to forecasting the weather, I believe that most people respond more intuitively to simple colours than to the complex units of data found in weather reports and downloadable apps. My phone can instantly inform me of the current temperature outside in degrees of Celsius, but this reading tells me nothing of how warm or cold it actually feels. How warm is 18°C, exactly? Does that mean I need a jumper or a coat? We can access a multitude of different kinds of data relating to the weather, but can this information be used to create something beautiful or intuitive to read?
Over the past couple of years, I have developed an increasing curiosity in the field of generative design, a process by which the designer creates a system, which is then fulfilled by an automated machine to produce one or many final outcomes. A designer might create a program that draws an artificial tree, changing the parameters of the tree with each iterative drawing to reflect the diversity that exists within nature. My earliest experiments were with simple digital drawing machines (while working through the excellent book Generative Design, available here. Each line could be drawn using very simple rules. For example, the lines follow the path of the cursor, or travel away from the cursor at changeable angles. As the rules become more complex, more sophisticated drawings can be created.
The weather itself is a system, a huge network of changing pressures and banks of cloud, constantly interacting with each other to create variations in temperature, precipitation, wind and humidity. I was intrigued by the prospect of revealing this system, particularly in a way that generated a visual product.
Joseph Mallord William Turner was a Romantic painter, known for his hazy watercolour landscapes. I think one particular strength of his paintings are the atmosphere they create. His applications of colour create very abstract impressions.
‘Study of Sunlight’, J.M.W Turner
Circa 1830, 190 mm x 190 mm
V&A Museum, UK
‘Holy Island, Northumberland’, J.M.W Turner
circa 1829, 431 mm x 292 mm
V&A Museum, UK
Images from the V&A website.
Turner seemed less concerned with painting specific places than with the dramatic possibilities of sea and air, and with the motion of the elements. In contrast to the objective scientific data presented by weather forecasting services, Turner’s paintings evoke an emotional response to the environment. Could I create a system that presented weather data with similar dramatic effect?
I was also interested in how generative approaches to graphic design could be applied in a practical way. Field are a design studio in London, specialising in interactive digital installations, such as their project ‘10,000 Digital Paintings’: “Invited to inspire the design community about the unlimited possibilities of digital print and generative design, FIELD created 10,000 unique cover artworks for the 2011 edition of paper manufacturer GF Smith’s legendary brochures. Each sleeve features a different view on a hypercomplex sculpture, generated through a process pairing generative coding with creative intuition. The energy of a dynamic process – caught in a timeless medium,” (from Field’s project description).
The digital paintings are based on random shapes. With each print, the viewpoint, structure and colour of the shape changes randomly. The rules dictating this were programmed into a piece of software that could create many more visual iterations of the shape than any human, meaning that each print is completely unique, echoing the premium message of the client. The strength of generative design is in creating automated variations of something: visuals that are dynamic and can react to certain parameters.
The visual identity for the Nordkyn tourist board, for example, changes to represent the current weather in Nordkyn, northern Norway. The creators, Neue, explain: “The visual identity is based on two main ingredients; our newly developed payoff, “Where nature rules,” and weather statistics from the Norwegian Meteorological Institute. A feed of weather statistics affects the logo to change when the direction of the wind or the temperature changes. On the website, the logo updates every five minutes. Nordkyn is truly a place where nature rules, even over the visual identity,” (Neue project description).
The Nordkyn mark is interesting because it rejects the traditional notion of a static logo, instead using a procedural mark that directly expresses weather conditions in a way that creates a unique visual brand. These approaches transform dull or incomprehensible data into something warm, engaging and fantastic.
Inspired by rock formations that are shaped and eroded by wind and rain, I explored creating physical sculptures generated from the weather conditions at a specific place. The sculptures digitally ‘grow’ as the day progresses, dynamically altering their shape to reflect changes in temperature, wind direction and the colour of the sky. In the same way that stalagmites are formed from the slow build up of limestone sediment in drops of water, the sculptures could be printed layer by layer, using 3D printing technology, providing a visual record of a day’s weather (this part of the project is still work-in-progress, will post an update soon).
How can data be used to tell a story? James Bridle’s project ‘A Ship Adrift’ takes a weather station atop a London hotel and uses measurements of wind to power an imaginary air-ship, piloted by a lost, mad artificial intelligence. The craft plots its fictional course across Google Maps, changing direction, slowing down and speeding up as the wind in London does the same, creating an adventure powered by both the wind and the imagination of the viewer.
Roni Horn’s photographic series ‘Still Water (The River Thames, for Example)’, images focuses on a small area of the surface of the river Thames. The colour and texture of these watery surfaces varies dramatically between images: colours range from black to blue and from dark green to khaki-yellow, and in each case the water’s texture is differently augmented by tidal movement and the play of light. By focusing on a single subject over a long period of time, Horn shows us the huge variations in the river’s appearance, changes that can sometimes be too slow for the human eye to recognise. These projects give us a new perspective on their subjects, allowing us to explore them in a way we would not be able to otherwise.
The images collected in this book are created using a generative system. Using a php script, images were downloaded from 64 public-access webcams across Europe. Some are amateur weather-stations, some are maintained by hotels or universities. Each recorded the sky every hour, taking 24 snapshots every day. The colour of the sky was extracted from these photographs as numeric values of red, green and blue, then plotted back to a geographic map, forming an image of the sky that extends beyond a single perspective. This is network vision, seeing through many lenses at once.
The webcams vary drastically in position, direction and the quality of the images they capture. Webcams facing East and West record the intense reds and yellows of sunrises and sunsets. Each webcam has a different sensitivity to light, each interpreting the blackness of night as various shades and hues of grey. Many cameras have uncalibrated white-balance settings, saving photographs that contain unexpected greens and yellows.
Often, London and Paris have an orangey-purple hue at night. Hannover is a bright white during night hours, the result of light refracting through a spider’s web infront of the camera, woven throughout the day. Svalbard is so far north that it never fully reaches darkness. A strange mix of expected and unexpected results.
Hampstead Heath, showing the glow of London at night
05.05.13, 20:42 GMT
Hannover, covered by spider web
01.05.13, 00:21 GMT
Amsterdam, a green sky created by odd white-balance
03.05.13, 09:43 GMT
The glitches and photographic mistakes become part of a painting process, where the cameras add their own idiosyncrasies to the image; a network of imperfect eyes.
Here’s a flick-through of the book, containing a sequence of images documenting the sky every hour for 7 days, from the 29th April to the 5th May 2013. If you click through to Vimeo, it’s available in HD.
This text is a modified version of the appendix in the aforementioned book. The book is on display at the Kingston University Degree Show, until 21st June 2013, and Work Out exhibition, London 27th-29th June 2013.
Many thanks to Richard Brauer, for an enormous amount of help with the coding of this project.
Thanks for reading!