Reef is a smart thermostat that understands your comfort and helps you save energy.
Reef is a smart thermostat that helps you save energy according to your comfort level. Through learning human comfort level, it creates a personalized energy saving plan.
Google Internet of Thing Research Award
UX Researcher
UX/UI Designer
UX MethodPersuasive des.
Affinity wall
Fast prototyping
User enactment
Software
Sketch
InVision
Illustrator
Photoshop
Reef is a thermo-comfort prediction system. It uses in-house sensors and wearable devices to collect thermo-related data and calculates personal comfort level based on the data. I worked with a team of three to design a futuristic thermal app that incorporates the information of personal comfort level and encourages people to save energy.
My role in this project including:
A primary function of a smart thermostat is to avoid unnecessary waste of energy. Whereas there are many existing smart thermostats, the personal comfort level is a new type of data that has never been explored. Little is known about how the representation of personal comfort level can contribute to a smart thermostat and help people save energy.
We sought inspirations from the literature including eco-feedback, persuasive design and information visualization. Using an affinity diagram, we summarized the ideas into three design dimensions of interest:
Based on the design dimensions, our team started to imagine the possible interactions that may influence people's thermal-related behaviors. Through sketching and wireframing, I generated stacks of ideas about incorporating personal comfort level in a smart thermostat. After five iterations, we converged our ideas into 15 interactions.
I created 30+ sketches to quickly explore the possible functions of Reef and used wireframes to specify the interaction steps.
People's energy consumption behavior varies in different environmental settings. It is hard for users to imagine their behaviors without providing them any context. To better probe users' reactions, we arranged a smart home environment with several grounding scenarios and observed the users’ reactions to our designs in each scenario.
The baby would be the most important. My wife and I could live with any temperature as long as the temperature in his room was comfortable.
If I’m cold, I would bump it probably only one degree and deal with it... my wife will get too hot, so I bump it at one degree.
Contrary to our expectation, norm-activation and rational thinking were not very effective in persuading users to change their thermal-related behaviors. Many users will neglect how much they can save and just increase the temperature when they were cold. However, the comfort level of other people in the house significantly affected one’s decision.
Users can see the the comfort levels of family members and adjust the temperature without being distracted by too much information.
Users can still know the temperatures in other rooms or saving of their current setting by tapping the bottom menu.
Proving individualized recommendations based on users' preferences and comfort data.
From this project, I learned how to explore a design field that I am not familiar with. I had to efficiently create a great deal of minimal viable prototypes to gather feedbacks. The discussion with the professors and ph.D. students allowed me to approach the design problem with a more comprehensive perspective. I also learned that how our preconceived assumptions could be different from the users' reactions. Another lesson I learned is the importance of simplicity in design. In our first round of design, we usually put too many details on one screen. Information overflow made users neglected those details. Therefore, how to present enough information without numbing users becomes the primary challenge of the design in the next phase.