Intent-Aware Data Visualization Recommendation

Abstract This paper proposes a visualization recommender system for tabular data given intents (e.g., “population trends in Italy” and “smartphone market share”). The proposed method predicts the most suitable type line, pie, or bar chart) visualized columns (columns used visualization) based on statistical features extracted from as well semantic derived intent. To predict appropriate type, we propose bi-directional attention (BiDA) model that identifies important table using intent parts of headers. determine columns, employ pre-trained neural language to encode both which are likely be visualization. Since there was no available dataset this task, created new consisting over 100 K tables their Experiments revealed our methods accurately predicted types columns.