Design explorations on establishing a feeling of responsibility for shared spaces.
M12 elective with Arthur Geel and Sark Xing.
Course: Data-enabled design. Mar. 2019 – Jun. 2019. TU/e
Shared housing is a common reality for young people worldwide. In this household model, two or more unrelated persons live together in a house where they share some faculties (i.e., bathrooms, kitchens). The dynamics of sharing facilities can create challenges in daily interactions: a common cause of frustration is facilities not being cared for well enough. In this project, we applied the Data-Enabled Design approach to guide our design process. Throughout the process, a strong emphasis was placed on gathering data to inform the design process. This project consists of two main stages: the Contextual Step and the Informed Step. In the first step, we sought to explore the design context by the use of data-gathering probes and qualitative questionnaires. In the second step, we aimed to apply the insights gained to work towards creating healthier living environments.
The Contextual Step
In order to create value through design, we need to have a thorough understanding of the design context and its actors. We deployed a series of devices that register sensor readings of its context. We included sensors that measured dust, humidity and temperature, and an electric fan to ensure the probe has the airflow required to make accurate measurements. These measurements were sent in 2-second intervals to the Data Foundry platform using OOCSI. In addition to the data probes, we explored the design context through qualitative data: self-reported questionnaires and stakeholder interviews. In interviews, we presented a number of data visualizations to users to get them to discuss what happened to cause the observed patterns in data.
The data probe and the detected data patterns
We gained three main insights during the entirety of the contextual step. Firstly, we found that the perception of air quality is positively correlated with cleanliness (including tidiness, smell, and other factors). Secondly, we found that shared spaces often have implicit rules, which direct the social dynamics. Finally, we found that these implicit rules are fragile: violations of rules lead to a loss of control in the shared context. We hypothesized that in order to create a healthier shared living environment, the inhabitants need to have a better overview of the accountability of individuals.
The Informed Step
In this step, we firstly developed a web application that kept track of how shared environments are used. Upon entering the shared room, a person can use the button containing their name to keep track of their presence. The presence of all users is displayed in a bar chart, so that everyone can see who has been in the kitchen. As a result, we did not see an increase in the feeling of responsibility after seeing others’ presence. Also, we found that the frequency of self-reported submission decreased over time. However, we also learned that our participants sometimes blame others on the Whatsapp group for messing up the kitchen. We see this blaming behavior as a way to remind the responsibility in the shared kitchen, which perhaps could be incorporated into the next iteration as a design opportunity.
An overview of iteration 1
Details of the iteration 1 prototype
For the final iteration, the web app prototype was adjusted to produce more direct value for the people of the shared house. In this app, the participants were able to see how much they contributed to taking care of the shared kitchen. The height of the lava layers visualized the blame given to the individuals for not taking care of the kitchen. As time progressed, the height of the layers was updated, potentially progressing towards an eruption. This experiment yielded mixed results. The overall concept of blame being visualized through the volcano was easy to understand for all participants, yet the way the blame was calculated was not transparent enough to be understood. Also, although we used blame as a motivator, sometimes the ones we blamed were not actually responsible for causing what they are accused of. Instead of blaming those that did not perform well, the system could share praise for those that did: being given a false compliment is certainly easier to handle than being falsely accused.
An overview of iteration 2
Stills of the eruption animation and the ending screen
In this 8-week elective, I learned the overall knowledge of data-enabled design from the lectures and the experience of design iterations. Firstly, I experienced how to use data in a qualitative way. It’s interesting to see that some visible data differences can motivate participants to recall some events that were unvalued to be helpful in reporting. Secondly, I learned in which condition the data-enabled design method can be applied. The method is constructive in gaining insights when users will use the prototype in their daily routine. It would be more valuable if the study is conducted in a field condition for a relatively long time for observing behavior changes. Finally, although I personally like the way of making something fake for quick insights and I also use the Wizard of Oz method many times in other user tests, I cannot agree that we can make too many fake things in the data-enabled design based user tests (or in a field study). As this method is totally in field condition, it gives the participants a lot of time to discover that they are interacting with a fake thing and then lose trust with the system.