Data2Vis: Automatic Generation of Data Visualizations Using Sequence-to-Sequence Recurrent Neural Networks

Abstract

Rapidly creating effective visualizations using expressive grammars is challenging for users who have limited time and limited skills in statistics and data visualization. Even high-level, dedicated visualization tools often require users to manually select among data attributes, decide which transformations to apply, and specify mappings between visual encoding variables and raw or transformed attributes. In this paper we introduce Data2Vis, an end-to-end trainable neural translation model for automatically generating visualizations from given datasets. We formulate visualization generation as a language translation problem where data specifications are mapped to visualization specifications in a declarative language (Vega-Lite). To this end, we train a multilayered attention-based encoder-decoder network with long short-term memory (LSTM) units on a corpus of visualization specifications. Qualitative results show that our model learns the vocabulary and syntax for a valid visualization specification, appropriate transformations (count, bins, mean) and how to use common data selection patterns that occur within data visualizations. Data2Vis generates visualizations that are comparable to manually-created visualizations in a fraction of the time, with potential to learn more complex visualization strategies at scale.

Towards Automated Infographic Design: Deep Learning-based Auto-Extraction of Extensible Timeline

“…DeepEye [33] applies ML models and design rules to determine whether a visualization is "good" or "bad" and recommends the "good" candidates. Data2Vis [14] uses a Recurrent Neural Network to automatically translate JSON-encoded datasets to Vega-lite [45] specifications. Draco [37] learns weights between hard and soft constraints that represent users' requirements and design guidelines.…”

VizNet

“…In a natural extension to these earlier ideas, researchers have recently introduced machine learning-based systems for automated visualization design. Data2Vis [18] uses a neural machine translation approach to create a sequenceto-sequence model that maps JSON-encoded datasets to Vegalite visualization specifications. Draco-Learn [49] learns tradeoffs between constraints in Draco.…”

“…Beginning with Cleveland and McGill's seminal work [11], researchers have studied this question of graphical perception by conducting human subjects experiments. And increasingly, researchers are seeking to operationalize the guidelines such studies produce using handcrafted rule-based systems [49,73] or learned models [18,27,41].…”

View recommendation for visual data exploration

“…Motivated by the need for an efficient and effective visual data exploration process, several solutions have been proposed towards automatically finding and recommending interesting data visualizations [6][7][8][9][10][11][12]. The main idea underlying those solutions is to automatically generate all possible visualizations, and recommend the top-k interesting visualizations, where an interestingness of a view is quantified according to some utility function.…”

“…Moreover, the hardest part of visualization recommendation is to quantify what would be interesting for the analyst. Consequently, the visualization recommendation problem has been approached from different angles, such as deviation-based/similarity-based methods that 1 Figure 1.1: Data exploration process-Key steps quantify interestingness as a similarity/distance metric [6,[13][14][15], user-actions-based methods that quantify interestingness in term of user's intent, which is inferred by her present actions or by historic data [7,16], perception-based methods, that learn human perception and use it to mine and recommend interesting visualizations [10][11][12].…”