A Factorial Analysis of Visual Profiles of Websites on Mobile Devices
(2024) International Journal of User-System Interaction — Vol. 16, n° 2, p. 25-48 (2024)
Files
IJUSI-16-2-Burny.pdf
Open Access - Adobe PDF
- 5.19 MB
Details
- Authors
- Abstract
- Evaluating a user interface often means comparing it against a reference, whether theoretical or empirical. To ensure an appropriate comparison, this paper contributes to computational evaluation by creating a large dataset (top 53 websites ranked in 20 categories: N=1060), computing 53 metrics for them (56,180 entries) by applying a principal component analysis to identify standout metrics, and by performing a factorial analysis to derive a reference visual profile for websites within each category. After analysing the correlations between these 53 metrics and performing various rotations, we have been able to reduce the expressiveness of these metrics to seven latent factors: colorfulness, color complexity, analogous color scheme dissimilarity, primary hue, lightness deviation, complementary colors dissimilarity, and structural simplicity. Therefore, positioning a screenshot with respect to the corresponding visual profile can be simplified based on these latent factors. This approach aims to provide a more comprehensive and context-specific understanding of computational evaluation in the realm of graphical user interface design.
- Affiliations
Citations
Burny, N., & Luis A. Leiva. (2024). A Factorial Analysis of Visual Profiles of Websites on Mobile Devices. International Journal of User-System Interaction, 16(2), 25-48. https://hdl.handle.net/2078.5/233481 (Original work published 2024)