Abstract
Existing approaches to data visualization are one of these two: accessible to end-user developers but limited in customizability, or inaccessible and expressive. For instance, commercial charting tools are easy to use, but support only predefined visualizations, while programmatic visualization tools support custom visualizations, but require advanced programming skills. We show that it is possible to combine the learnability of charting tools and the expressiveness of visualization tools. Uvis is an interactive visualization and user interface design tool that targets end-user developers with skills comparable to spreadsheet formulas. With Uvis, designers drag and drop visual objects, set visual properties to formulas, and see the result immediately. The formulas are declarative and similar to spreadsheet formulas. The formulas compute the property values and can refer to data from database, visual objects, and end-user input. To substantiate our claim, we compared Uvis with popular visualization tools. Further, we conducted usability studies that test the ability of designers to customize visualizations with our approach. Our results show that end-user developers can learn the basics of Uvis relatively fast.
Highlights
Data visualization aims at supporting human abilities by showing data using visual variables such as position, color, size, and orientation [1]
Charting tools allow designers to build a visualization by selecting predefined visualization templates and mapping them to data
Like charting tools, there is no way to create visualizations beyond what is predefined. Visualization tools such as Lyra [5], iVisDesigner [6], and Data Illustrator [24] allow the creation of custom visualizations without real programming
Summary
Data visualization aims at supporting human abilities by showing data using visual variables such as position, color, size, and orientation [1]. Designers can adjust the visual properties to a limited extent This approach improves learnability (learning how to build a visualization) as well as task efficiency (building it fast). It lacks expressiveness (customizing the visualization to specific needs). Data analysis tools such as Tableau and its predecessor Polaris [2] integrate well with existing data and help users explore the data. Like charting tools, there is no way to create visualizations beyond what is predefined Visualization tools such as Lyra [5], iVisDesigner [6], and Data Illustrator [24] allow the creation of custom visualizations without real programming.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
