Measurement Of Information Aesthetics In Time Series Data Visualization

Time series data is an important kind of structured data,the representation of which using visualization technologies can help users explore it more intuitively and find its underlying laws.At present,most of the research on time series data visualization focuses on the optimization of visualization technology,while ignoring the study of visualization aesthetics.In fact,a good data visualization requires a balance between functionality and aesthetics.To achieve this balance,we need to measure the aesthetics in the visualization of time series data at first.Based on the information aesthetic model in the context of visualization proposed by Lau A and Moere A V(2007),a set of metrics are created to measure the information aesthetics in time series data visualization.The creation process is divided into five steps.First,candidate questions were obtained through semi-structured interviews(n = 15).Then,through expert classification,these candidate questions were structured according to the three design dimensions of information aesthetics,namely data,interaction,and aesthetics(n = 3).Second,scorers familiar with visual design assessed the surface validity of candidate questions(n = 8).Third,after preliminary revision,the candidate questions were applied to evaluate the actual time series data visualization works(n = 84).Through reliability analysis,exploratory factor analysis,and confirmatory factor analysis,the initial design dimensions and their cooresponding problems were explored and improved.Fourth,the initial scale generated in the third step was regarded as a guide for design practice.Fifth,the results of design practice were used to conduct a second round of evaluation(n = 120).Using reliability analysis,confirmatory factor analysis,and target association validity analysis,the stability,accuracy,and differentiation of the questionnaire were measured to further strengthen and improve the questionnaire.After the above five steps,a set of scales based on three information aesthetic dimensions and 17 questions was finally obtained.The scale can be used to measure the degree of information aesthetics in the visualization of time series data,and can also be used to guide the visual design of time series data to achieve a balance between functionality and aesthetics.