The course Making Sense of Sensors is designed to teach students how to work with large datasets. To do this in a realistic setting, we worked in a group of four students to collect data from participants for answering our research question:
To what extent are people able to estimate their daily step count, and does that differ between active and inactive days?
Four office workers participated in this study. The required apps for saving the amount of steps measured by the MiBand3 were installed by the researchers on the phones of the participants, and an informed consent and instructions were given. To comply with the ethical principles, participants were fully informed and safe and were told that they could stop participating at any time without explanation. The participants wore the MiBand3 for two consecutive weeks and were daily reminded to fill out questionnaires for each day. The participants were asked to estimate the amount of steps they had taken, both for the day in total and per time of day (categories: before noon, between noon and five, after five). They were also asked whether they had taken part in any activities, whether they had gone to work, and whether they would describe the day as active, moderate, or inactive. To ensure anonymity, participants were given a number which they also had to fill out in the questionnaire.
To make the acquired data findable, we gave them unique and persistent identifiers that are specified by meta data. Data are accessible via Python, a standardized, open, free, and universally implementable communications protocol. By using CSV format, we made the data interoperable. Finally, the data are reusable because of the well described meta data and provenance.
The current course has helped me obtain a better understanding analysing data in Python. Although I would rather use IBM SPSS since the interface allows for recognition instead of recall, I do to acknowledge the endless possibilities of Python that make it ideal for almost any type of data analyses. Being able to do different kinds of data analyses allows for a wider range of possible research designs for, for instance, user testing, which in turns supports more usable designs. That, together with the acquired knowledge on the use of sensors, gives the present course a high value for designers.