Home

Daan
Weijers

e-mail me

Nerviz: A new approach to self-reporting stress levels

A project with Philips Design — 2015

Image of the circa alarm clock on a nightstand

In short

The situation

As a group of four, together with Philips Design, we were asked to create a product or service with data at its core. We were given a large data set of student’s health, and no further constraints.

The outcome

Through a very iterative process, we first created a service that allows students to see if any places on their university campus would provide for a calmer environment to study, using indicated stress-levels and locations. Then, we reconsidered the qualitative data and proposed a new way of collecting stress levels through minigames, which resulted in more quantitative data.

My role

This part-time project on the Eindhoven University of Technology lasted for about three weeks. As a group we were doing the conceptualisation and design of the service and data gatherer. Personally, I was responsible for programming many of the initial visualisations in the app, as well as prototyping some of the minigames.

The data-gathering mini games we designed for the project

Getting the data

The dataset we were given by Philips was a very extensive record of student’s health on a university campus. It was especially comprehensive in the in the type of ‘measured’ data, such as heart-rate, weight, amount of steps taken in a day, etc. We also quickly noticed that the data set was lacking quite a bit of information in the stress column.

On the one hand, the feeling of stress is very personal, on the other hand, it is also often influenced by external factors. We felt that the 1–5 scale of stress level in the set therefore lacked a lot of context, and since stress (and mental health in general) can be critical for students’ performance, social lives, etc, we felt there was some opportunity to dig deeper.

Iteration 1

Every entry in the dataset was accompanied by a GPS location. We decided to place every entry that included a stress level on a map as a dot in processing, with the color corresponding to the stress level.There were lots of clusters of entries and some single entries.

All individual stress reports plotted to a map as coloured dots, scale: being 1-blue and 5-red

This visualisation & clustering of information sparked the idea of providing a service that could inform students on what location on the university campus would provide a calm or stressful environment, so they could use that knowledge to their advantage.

A new colour scheme we used after the first prototype, to increase contrast

First prototype

After the first week, we presented the first prototype, built around the initial data visualisation. It allowed people to explore the campus using a map that shows (clustered) reports of stress by location.

Iteration 2

In the first iteration, we did not take time into account, but just showed the tendency of a place to have a stressed atmosphere or not. Of course, a place can be very tense in the morning and super chill in the afternoon. To be usable as a product for students, this third dimension of time needs to be added.

Visualisation of occurring stress levels on one day.

We started by grouping the reports at a particular hour by stress level, to get both an insight on the total amount of reports at a certain time and the distribution of stress over time.

Admittedly, a circle was not the best way to compare different moments throughout the day, but we quickly saw that at some moments in the day, people were calmer than at other moments.

Second prototype

To make our product useable for students as a planning tool, we needed to show the reported stress levels of locations on certain times of the day. Next to that, we wanted to give our users an indication of the patterns in their own stress levels, to make them more mindful of that.

Data collection

Because the provided stress-data was based off self-reported stress levels, rather than measurements, the data was rather qualitative, and in contrast with the rest of the dataset that consisted of purely quantitative measurements. Together with Philips’ researchers, we discussed how interesting it would be to compare self-reported results to actual measurements.

Literature research

Through literature research, we discovered that being stressed basically influences three measurable factors:

  1. Perception of time.
  2. Fine motor skills.
  3. Gross motor skills.

The team thought it could reverse that effect, and measure stress by measuring the capability of the user to perform precision tasks.

Researchkit

We were inspired by Apple’s researchkit, where small and easy tasks are performed to provide data to health-researchers.

Minigames

During the remaining time, the team thought of and prototyped four minigames that would quickly be able to test the user on one of the three measurable factors of stress (perception of time, fine motor skills, gross motor skills).

Regardless of the personal perception of stress, the performance on these tests would provide quantitative data regarding a person’s stress level. For research purposes, we thought it would be interesting to compare the performance on the games to a self-reported stress level, which is why we designed a less numbers-based and more natural way of reporting stress.

Final iteration

Once in a while, a person would be asked to report their stress level through a minigame and the reporting function. In other cases, the app serves as a way to provide context to the data on campus, using the (refined) results of the earlier iterations.