Introduction

For this study, you will see a text description generated for an UpSet plot. An UpSet plot is a set visualization technique similar to Venn diagrams, but unlike Venn diagrams, UpSet works for more than three sets.

Our research aims to make data visualizations more accessible to people with visual impairments. We want to understand whether text can convey similar amounts of information as a chart. We first will introduce what an UpSet plot is, and how to interpret data from the plot.

UpSet Explained

UpSet plots the intersections of sets in a table. Each column corresponds to a set. Bars at the top of the columns show the size of the sets. The row corresponds to an intersection: marks in the cells show which set is included in the intersection. The number of sets that participate in the intersection is referred to as degree. If there is no mark in any of the cells, then it is the intersection of no set, which is also referred to as the empty intersection, with a degree of 0. If there is a mark in every cell of a row, then it is the intersection of all sets.

UpSet plots the size of the intersections as bar charts to the right of the table. The table is also useful because it can be sorted in various ways. A common way is to sort by size, but it's also possible to sort by degree or sets.

Imagine an UpSet plot that shows movie data. Movies have genres like Drama, Comedy, Thriller, Mystery, or Crime. A movie can have a single genre, or it can have multiple genres. In this example, the genres are the sets. Some sets are bigger: there are more Drama movies than Mystery movies, for example. And some intersections will be more common: Thriller and Mystery might be a popular combination, while the combination of Drama, Comedy, and Thriller might be rare.

Glossary of Terms

Here are some terms we use in the text description:

• Movies that don’t fall into any genre are an intersection of no set/the empty intersection.
• A row with only one mark for drama (movies that are just dramas and have no other genre) is an independent set intersection.
• A row with 2-3 marks like "Drama-Comedy" corresponds to a low-degree set intersection.
• A row with 3-5 marks (e.g., Mystery-Crime-Thriller) corresponds to a medium-degree set intersection. A row with even more marks is a high-order set intersection.
• The last case is that set containing all sets (i.e. all movie genres), which we call an all-set intersection.

That’s it for the introduction! Please stop now and ask the interviewers if you might have any clarifying questions.

Next, we’ll explore text descriptions generated for a variety of UpSet Plots. Click on the link below to see:

UpSet Plot Description