<i>KG4Vis:</i> A Knowledge Graph-Based Approach for Visualization Recommendation

Li, Haotian; Wang, Yong; Zhang, Songheng; Song, Yangqiu; Qu, Huamin

doi:10.1109/tvcg.2021.3114863

Cited by 67 publications

(32 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…According to our survey, no existing ML-based approaches can adaptively recommend multiple appropriate visualizations for a dataset. Existing ML-based approaches often recommend the only one visualization choice [1], [8], [10], [20], not adaptive to different datasets. Second, most ML-based approaches (e.g., Data2Vis [8] and VizML [10]) are built upon deep neural networks.…”

Section: A B Cmentioning

confidence: 99%

“…These approaches work as a black box and can undermine users' trust in the visualization recommendation results, especially for general users without a background in visualization and deep neural networks. A recent study, KG4Vis [20], presents a knowledge-graph-based approach to recommend visualization in an explainable manner for tabular datasets. But their explanations are built on single-column features and can only indicate which single-column features account more for a visualization type, which is called global interpretation [21].…”

Section: A B Cmentioning

confidence: 99%

See 1 more Smart Citation

Neighborhood Constraints Based Bayesian Personalized Ranking for Explainable Recommendation

Zhang

Zhu

Wang

2023

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Automated visualization recommendation facilitates the rapid creation of effective visualizations, which is especially beneficial for users with limited time and limited knowledge of data visualization. There is an increasing trend in leveraging machine learning (ML) techniques to achieve an end-to-end visualization recommendation. However, existing ML-based approaches implicitly assume that there is only one appropriate visualization for a specific dataset, which is often not true for real applications. Also, they often work like a black box, and are difficult for users to understand the reasons for recommending specific visualizations. To fill the research gap, we propose AdaVis, an adaptive and explainable approach to recommend one or multiple appropriate visualizations for a tabular dataset. It leverages a box embedding-based knowledge graph to well model the possible one-to-many mapping relations among different entities (i.e., data features, dataset columns, datasets, and visualization choices). The embeddings of the entities and relations can be learned from dataset-visualization pairs. Also, AdaVis incorporates the attention mechanism into the inference framework. Attention can indicate the relative importance of data features for a dataset and provide fine-grained explainability. Our extensive evaluations through quantitative metric evaluations, case studies, and user interviews demonstrate the effectiveness of AdaVis.

show abstract

Section: A B Cmentioning

confidence: 99%

Section: A B Cmentioning

confidence: 99%

Neighborhood Constraints Based Bayesian Personalized Ranking for Explainable Recommendation

Zhang

Zhu

Wang

2023

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…Specifically, the knowledge can be standard procedures [21] and linguistic rules [48] collected from domain literature, relationships between samples [16,51,74] specified by users, constraints distilled from expert experiences [46], numeric features calculated based on pre-collected samples [72,73], etc. Besides, many recent works attempt to utilize knowledge involved in off-the-shelf digital resources, such as ontology [41,59], corpus [82], knowledge graphs [8,37], pre-trained models (e.g., knowledge distillation) [75], etc. There have been literature reviews on techniques for integrating human knowledge into machine learning models [17,71], and many of them are also applicable in visualization.…”

Section: Knowledge-assisted Visual Analyticsmentioning

confidence: 99%

Incorporation of Human Knowledge Into Data Embeddings to Improve Pattern Significance and Interpretability

Zhou

2022

IEEE Trans. Visual. Comput. Graphics

View full text Add to dashboard Cite

2 a1 b1 c1 b2 a2 c1 c2 Fig. 1. The visualization system interface that consists of three components: (a) Knowledge Editor (KE), (b) Data Projector (DP), and (c) Pattern Explainer (PE). The analyst is analyzing the Covid-19 dataset [2] that consists of 173 sequences, each recording the number of monthly-confirmed cases (per million people) from May 2020 to April 2021 of a country. The analyst creates six country groups (see colored tree nodes) according to their knowledge using (a) and incorporates the class information into the embeddings of countries to generate the projection with clear and class-consistent visual structures that foreshadow underlying data patterns. (b2)The embedding projection without knowledge where does not exist visual structure.

show abstract

“…Thus, by using the links to visualizations provided in VizML [18], we built a crawler to collect the SVG-bitmap visualization pairs from Plotly Chart Studio 7 . Following previous practices using the VizML corpus [18,22], we also kept one visualization per user. Also, we removed those pairs with invalid bitmaps or SVGs, for example, empty visualizations or incomplete SVGs.…”

Section: Corpusmentioning

confidence: 99%

Structure-aware Visualization Retrieval

Li,

Wang,

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

With the wide usage of data visualizations, a huge number of Scalable Vector Graphic (SVG)-based visualizations have been created and shared online. Accordingly, there has been an increasing interest in exploring how to retrieve perceptually similar visualizations from a large corpus, since it can benefit various downstream applications such as visualization recommendation. Existing methods mainly focus on the visual appearance of visualizations by regarding them as bitmap images. However, the structural information intrinsically existing in SVG-based visualizations is ignored. Such structural information can delineate the spatial and hierarchical relationship among visual elements, and characterize visualizations thoroughly from a new perspective. This paper presents a structureaware method to advance the performance of visualization retrieval by collectively considering both the visual and structural information. We extensively evaluated our approach through quantitative comparisons, a user study and case studies. The results demonstrate the effectiveness of our approach and its advantages over existing methods.

show abstract

KG4Vis: A Knowledge Graph-Based Approach for Visualization Recommendation

Cited by 67 publications

References 43 publications

Neighborhood Constraints Based Bayesian Personalized Ranking for Explainable Recommendation

Neighborhood Constraints Based Bayesian Personalized Ranking for Explainable Recommendation

Incorporation of Human Knowledge Into Data Embeddings to Improve Pattern Significance and Interpretability

Structure-aware Visualization Retrieval

Contact Info

Product

Resources

About