Interactive visualization method for exploratory data analysis

We propose an interactive visualization method suitable for exploratory data analysis. In this type of data analysis, statistical charts are employed to help the user understand the target data, and a displayed chart is redrawn interactively according to the user's input inspired by looking at the chart. Two factors must be considered if the chart is to be suitably redrawn: the type of chart displayed and the parts of the chart that are changed by the user's input. We developed our proposed method based on these factors. As the first step of the method, the part to change and the details to meet the given user's input for the displayed chart are interpreted. Then, a semantic frame used to generate a statistical chart from data is updated by the results of the interpretation, and several candidates as appropriate chart types tire chosen based on the frame and chart-type determination rules. Finally, the most preferable chart is chosen from among the candidates according to the type of user input, and the displayed chart is transformed into a new chart with animation. BY using this method, data can be visualized interactively, according to the change of the user's viewpoint without interrupting the thinking process.