# Calliope: Automatic Visual Data Story Generation from a Spreadsheet

**Authors:** Danqing Shi, Xinyue Xu, Fuling Sun, Yang Shi, Nan Cao  
**arXiv:** [2010.09975](https://arxiv.org/abs/2010.09975)  
**Date:** 2020-10-20  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2010.09975v1/

## Key points

- Visual data stories are widely used in data-oriented storytelling to enhance comprehension and retention of story content.
- Calliope is a novel system that automates the generation of visual data stories from an input spreadsheet.
- Calliope incorporates a logic-oriented Monte Carlo tree search algorithm to explore the dataset and generate relevant "data facts."
- The importance of each data fact is determined using information theory.
- Calliope visualizes each data fact using charts and provides automatically generated descriptions for better understanding.
- The system includes an online story editor for easy revision of the generated story.
- The proposed technique is evaluated through three example stories, two controlled experiments, and interviews with 10 domain experts.
- The evaluation demonstrates that Calliope significantly improves efficiency in visual data story generation.
- Calliope simplifies the process of creating visual data stories from spreadsheets.
- Calliope has great potential for facilitating efficient and effective storytelling with data.

## Summary

In the field of data-oriented storytelling, visual data stories presented in the form of narrative visualizations like posters or data videos are widely used to enhance comprehension and retention of story content. To address the challenge of generating these visual data stories due to technical barriers such as data analysis, visualization, and scripting, this paper introduces a novel system called Calliope that automates the generation of visual data stories from an input spreadsheet. Calliope incorporates a logic-oriented Monte Carlo tree search algorithm that explores the dataset provided by the spreadsheet. This algorithm progressively generates "data facts," which are pieces of information relevant to the story, and organizes them logically. The importance of each data fact is determined using information theory. Furthermore, Calliope visualizes each data fact using charts and provides automatically generated descriptions for better understanding. The system also includes an online story editor that allows easy revision of the generated story. The proposed technique is evaluated through three example stories, two controlled experiments, and interviews with 10 domain experts. The evaluation demonstrates that Calliope significantly improves efficiency in visual data story generation. Overall, this paper presents a valuable contribution to the field by introducing an automated system that simplifies the process of creating visual data stories from spreadsheets. The results indicate that Calliope has great potential for facilitating efficient and effective storytelling with data.