Visualization Study of the Relationships between Responses in Choice-Type Questionnaire

Questionnaire is often used to investigate the respondent´ s opinion toward some product or service. To realize the knowledge mining from response data of questionnaire, this paper proposes one kind of visualization technique based on frequency analysis together with co-occurrence analysis. In created visualization graph, the nodes express the result of frequency analysis and the links between nodes show the result of co-occurrence analysis. Different from traditional co-occurrence analysis in which single case of the respondent´ s attribute such as age and gender is focused on, proposed technique allows to integrate the information of two graphs about two cases of an attribute in single graph. Appling this technique to an example dataset including two attributes and five questions, this paper discusses the applicability and limitation of this visualization technique.