Math, Data & Computing

Where I started …

Throughout my bachelor’s, I developed skills in statistical analysis and visualizing these in plots using Python. By attending the Data Analytics course, I was able to do this with given datasets. However, I felt less confident in performing these methods on my own datasets and being sure the outcomes I yield are correct.

Regarding qualitative data gathering and analysis, I enjoy the human side of the data retrieved from interviews. So, I spent the B3.1 internship expanding my skills to analyze and visualize that.

  • Posters Data Analytics course: here and here
  • Poster Making Sense of Sensors: here , and group report: here

… how I continued …

By becoming interested in the perception of emerging technologies and our hesitations with them, I entered a complex domain of observing uncommon interactions between people and technology. Primarily, I get a grip on this complexity by gathering qualitative interview data and performing thematic analysis. Additionally, I employ quantitative methods where applicable to find relations between variables. For example, in the Data-enabled Design course, I compared sleep quality in relation to behaviour.

I have touched briefly upon neural network design in the Embodying Intelligent Behavior course. Although I see the value of personalization of products, I do not feel enough enthusiasm for it to continue its usage in my further work.

  • Report Data-enabled Design: here
  • Report Embodying Intelligent Behavior: here

… and where I’m headed.

I was not confident in determining what quantitative data to collect and how to properly analyze it to gain insights. The CWI collaboration taught me much about data collection during a study. From small things like asking whether a participant has VR experience when doing a VR study, to deciding upon meaningful post-trial questions and data logging. For example, logging interpersonal distance combined with the Likert-scale question on how much distance a participant wanted to keep.

Besides collecting data, analysis determines the answer to research questions. For numerical data, like the logged distance and Likert-scale questions, I used the ILLMO software and instruction videos to perform basic statistics to determine significant differences between conditions. For me this was already quite an achievement. For future development to create publishable research, I should practice doing these analyses to gain confidence.

Lastly, working with the PID algorithm taught me to be critical when implementing an algorithm to do computing work. After extensive trial-and-error to control the algorithm, no suitable computing behaviour was found. It did teach me though how the algorithm should work and I was able to replicate this in my own behaviour script.

  • M2.2 report, including appendix on which pre- and post-study data was gathered: here
%d bloggers like this: