Final Thesis

Data analytics of urban soundscape perception

In the research project “Person-focused Analysis of Architectural Design”, we study human perception of built environments. More specifically, we address the visual and auditory experience of virtual or virtually reconstructed urban spaces. The soundscape evaluation in such spaces contains ratings and individual vocabularies, which highly depend on the cultural background of people, among others. This dataset is already collected for the IHTApark environment (photo left).

The task for this thesis is to investigate and develop statistical methods which are able to unveil the most salient aspects of individual perception and subject’s clusters in the dataset. The methods aimed to be developed are analysis of variance (ANOVA), principal component analysis (PCA), confidence ellipses, and multiple factor analysis (MFA). For this reason, a first literature overview of the existing methods will be required. After this, the core task will consist on the implementation of such methods using the dataset, to extract conclusions on how people’s perception is affected by the soundscapes.

Requirements
- Matlab skills
- Statistics skills
- Accurate and reliable work attitude