Glyphs for exploring crowd-sourced subjective survey classification

Kachkaev, A., Wood, J. & Dykes, J. (2014). Glyphs for exploring crowd-sourced subjective survey classification. Computer Graphics Forum, 33(3),

[img]
Preview
PDF
Download (2175kB) | Preview

    Abstract

    The findings drawn from opinion survey responses are usually made by producing summary charts or conducting statistical analysis. Both involve data aggregation and filtering as exploring the unaggregated data has traditionally been impractical or error-prone for large numbers of responses. We propose the use of glyphs with parallel coordinate plots to show all survey responses in a single view and design an interactive visual analytics tool around the representation to explore the data. We use this software for a ‘photo content assessment’ survey, where 359 participants classify 900 images by seven criteria. The proposed approach allows all 8,434 responses (49,285 answers to questions in total) to be represented in a single view and helps analysts to both clean the data and understand the nature of the survey responses. We describe the construction of the survey response glyphs and the interface to the interactive visual analytics software and generalise the design principles that arise from the approach. We apply the tool to two other datasets to evaluate the technique and to confirm its wider applicability for surveys with Likert scale responses.

    Item Type: Article
    Additional Information: The definitive version is available at http://diglib.eg.org/ and http://onlinelibrary.wiley.com/.
    Subjects: Z Bibliography. Library Science. Information Resources > Z665 Library Science. Information Science
    Divisions: School of Informatics > giCentre
    URI: http://openaccess.city.ac.uk/id/eprint/3393

    Actions (login required)

    View Item

    Document Downloads

    More statistics for this item...