We created "Nerviz", a new approach to self-reporting stress data
This project in 60 seconds
For the "Beautiful Data Products" course we worked together with the team at Philips Design. We were asked to create a product or service with data at its core. The initial assignment was to explore and create visualizations based on a large data set containing various parameters about the health of students at an American college. We found a data set containing self-reported stress-levels with a timestamp and a location and used those to visualize when and where students are most stressed.
After our initial experiments, we became interested in the collection of mental health data. Based on literature our hypothesis was that how people interact with their mobile phones could say something about how they feel and how stressed they are. We created an experimental set of minigames to test people's fine motor skills, gross motor skills, and timing to see if they relate to their self-reported stress levels.
I created the initial prototypes of self-reported stress levels on the map and clock using Processing. For the final assignment, I worked on the concept and wireframes of the mini-games.
For the initial assignment, we got an open data set with a variety of data points about the health of a group of college students. Our team decided to take a hands-on approach to see what could be done with the data in the dataset, creating several smaller prototypes during the process.
We found a set of self-reported stress levels including a location and timestamp and decided to start plotting these on a map of the campus to find the most stressed and relaxed locations on campus. Blue indicates a calm state while red indicates stress.
Although there are locations on the campus that show more blue or red, it doesn't give a lot of context to the data. A location could be very chill in the morning and changing to very tense later in the day. The objective of the course was to experiment so we decided to create a separate visualization of just time. We started by grouping the reports at a particular hour by stress level, to get both an insight into the total amount of reports at a certain time and the distribution of stress over time.
Admittedly, a circle is not the best way to compare different moments throughout the day. Visualizing 24 hours of data in a circle becomes confusing due to the resemblance of the 12-hour clock. We decided to split the data in AM and PM but a grouped bar chart would have been better.
To make it more useful for people on the campus to find a calm location we decided to make a timeline view where locations are grouped. It shows a timeline for your most used locations, the size indicates the number of reports and is an indicator of how crowded a place is, while the color shows the average self-reported stress level. On the right, we wanted to give our users an indication of the patterns in their own stress levels, to make them more mindful of that.
After presenting these first explorations we saw an opportunity to focus more on data collection. Wearables and smartphone apps have made it easier to gain data about daily activity or sleep. But at this moment you can't reliably sense indicators of people's mental state. You have to ask people to self-report during the day which has a higher threshold.
Through literature research, we discovered three measurable factors that could indicate stress:
We were inspired by Apple’s researchkit, where small and easy tasks are performed to provide data to health-researchers.
We decided to create an experimental prototype that would gather both smartphone interaction data and self-reported stress levels. These datasets could be used in a study to find a potential correlation between the two. We created a fun and light app with various mini-games to test timing, precision tasks and larger scale movements like balancing while walking.
Once in a while, a person would be prompted using a notification to report their stress level through a minigame and the reporting function. Our focus in the design was to make the task of entering data as simple as possible to get better results. When the user opens the app when there is no game to play it provides context to the data on campus, using the (refined) results of the earlier iterations.
I hope you enjoyed this project.
How about looking at another one below?