This Physikit project is an exploration into making data tangible and physical. Data is typically the domain of analysts and big data companies, and the way data is often presented – in time series or graphical form – can be challenging for a lay person to interpret. We wanted to explore if, by turning data into physical forms, we could make it more accessible and decipherable to a non-expert.
Physikit is a toolkit that enables anyone to quickly and easily connect data to physical visualisation cubes, thereby creating their own physical visualisations of data. The toolkit consists of two main parts. Firstly, there are four physical visualisation cubes – called physicubes, and secondly there is a very simple web interface through which anyone can make and manage connections between data and the physicubes. The four physicubes each have different physical properties and can represent data in different physical ways, through light, vibration, rotation, and airflow.
The light cube glows in different ways to shows how data is increasing and decreasing, or to alert when the data has gone above or below some threshold value.
The vibration cube vibrates in different ways to shows how data is increasing and decreasing, or to alert when the data has gone above or below some threshold value.
The rotation cube spins a star shaped disc in different ways to shows how data is increasing and decreasing, or to alert when the data has gone above or below some threshold value.
The air cube flows air in different ways to shows how data is increasing and decreasing, or to alert when the data has gone above or below some threshold value.
The data to connect to the physicubes could come from any data source, such as open data streams found online, or social media data, but in this project we decided to use the Smart Citizen Kit to provide streams of environmental data including temperature, humidity, light, sound, NO2 and CO.
We asked 5 different households across London to install a Smart Citizen kit and a Physikit in their homes for 3 weeks, to see how they would use the Physicubes and what they would learn about the environment within their home. In some households individuals took ownership of a single cube and would position it in their own bedroom, whereas in other households the cubes were more communal and positioned in rooms that everyone used. As the 3 weeks progressed patterns of use began to emerge with many people linking the light cube with the light sensor, or the air cube with one of the air quality sensors. The vibration cube was most commonly set to alert mode as it was best suited to creating a loud, noticeable alert, while the light and rotation cubes were more commonly used to show continuous changes in the data in the more ambient way that they best supported.
Some households incorporated the cubes into their surroundings by using them as candle holders, while one household put a basil plant on the rotation cube and placed it beside the window. They connected the cube to the humidity sensor and configured it to rotate only when the humidity was below a certain level. When they returned home after work, they could then see if the cube had rotated or not by looking to see if the plant was growing towards the window – indicating no rotation and high humidity, or if the plant was growing straight – indicating rotation and low humidity. As such they were able to create a history by adding items to their cubes.
In another household the mother tried to use the light cube linked to the sound sensor, to show her children how noisy they were. In fact, the cube showed that the mother was noisier than the children! Over the course of the 3 week study, the households became familiar with the expected behaviour of each cube and quickly noticed and investigated when the cubes acted in an unfamiliar way. By the end of the trial each household was convinced that the air quality sensors were faulty as nobody could work out the patterns of behaviour when the cubes were connected to these sensors. In fact, the households were correct – most air quality sensors require regular calibration to provide accurate results and since calibration was not performed on these sensors they were providing inaccurate data.
It is interesting to note that even the simple physical visualisations of data, afforded by the cubes, still led to substantial insights into the sensor data. Most notably, leading people to question the accuracy of the sensor technology itself which can often go unchallenged.
Check out the project website: