top of page
Doug Caldwell

Book Review: Visualizing with Text, by Richard Brath


Visualizing with Text (2020), by Richard Brath (Partner at Uncharted Software, Inc.) is an intriguing addition to the data visualization literature. While most data visualization books gloss over the topic of text, Brath’s fertile mind fully embraces the subject. His clever use of text can improve data visualizations and his use of visualization techniques can enhance traditional text. His cross-disciplinary approach offers a fresh perspective that expands the range of design possibilities.


The use of text across disciplines

Brath introduces the topic by showing the breadth of his vision with text examples in the areas of: cartography, typography, tables, science classification and notation, code editors, alphanumeric charts, art and poetry, graphic design and advertising, comics, postmodern text, and data visualization. It is simply not possible to view these examples without thinking of new ways to incorporate text in design.

Brath highlights the advantages of encoding data with text. These include fast decoding, reduced cognitive load, and easy recognition. As an example, use of a legend on a chart requires the reader to go back and forth between the graphic and the legend to understand the chart. When it is possible to directly incorporate the legend into the graphic, the cognitive load on the reader is reduced.


A framework for using text

Brath identifies the differences between text that is read and graphics. Reading text requires focus to understand a sequential string of words. Visualization leverages our pre-attentive capabilities, where we perceive patterns before consciously focusing. The author bridges the gap between the two modes with a unifying framework.

Like most symbols in data visualization, text can be displayed using the visual variables, such as position, size, shape, color, or brightness. Text greatly expands this range of options. It has the additional elements of typeface, weight, case, oblique angle (italic), underlines, width, x-height, serifs, stress, contrast, and angle. In addition, there are non-type visual attributes like background color, gradients, superimposition, outline, and drop shadows.

The concept of text scope further expands the design possibilities. Brath defines text scope as a spectrum from characters, to syllables, words, phrases, sentences, paragraphs, and documents. We could also add collections of documents to this list. He finishes his framework with a discussion of three text layouts: prose, tables, and lists. The combination of visual variables, text variables, text non-type visual attributes, scope, and text layout form the foundation for his ideas on visualizing text.


Text in Data Visualizations

Brath moves methodically through a series of examples showing how text can be incorporated with data visualizations. He starts with the introduction of coded and full text labels as point marks. The advantage of this approach is that the text is frequently identifiable without resorting to a legend or tooltips.

His example of a scatterplot of National Parks shows how park names can be included in a visualization that also shows the number of visitors, park size, age, and region. Each park name is easily identifiable and integrated with other information. When situations become more complicated, Brath identifies techniques to accommodate difficult labeling cases, like large numbers of labels or very long labels.


Scatterplot of U.S. National Parks showing multiple variables.
Scatterplot of U.S. National Parks showing multiple variables.

Brath introduces his discussion of distributions with stem and leaf plots, using examples from rail timetables, poverty by state, and stock market sectors. He modifies traditional stem and leaf plots by using text, rather than numbers, to expand the leaves.

In an example of U.S. population density versus gun murders, he incorporates two variables in the text using font weight and color, while using the two-letter state abbreviations to identify the states. He makes this possible by rounding the poverty rates to the nearest percent.


Stem and Leaf Plot of U.S. state poverty, population density and gun murders.
Stem and Leaf Plot of U.S. state poverty, population density and gun murders.

Another variation of the stem and leaf plot shows the adjectives associated with characters in the Grimms’ fairy tales. The stem lists the characters, while the leaves show adjectives describing the characters, with adjective frequency decreasing from left to right with decreasing text boldness.


Stem and leaf plot of characters and adjectives describing them.
Stem and leaf plot of characters and adjectives describing them.

Brath discusses more complex variants of stem and leaf plots before introducing directed graphs which show a root phrase, followed by other phrases that follow the root phrase and subsequent phrases. The image below shows phrases from Alice in Wonderland.



Directed graph of root phrase.
Directed graph of root phrase.

Brath moves on from points to microtext lines, where the text is used to describe the lines. He identifies three options, 1) where text completely replaces the line work, 2) where text is integrated with the line work, and 3) where the line work is dominates and is and labeled with adjacent text. Brath is exhaustive in exploring his options and some choices are more effective than others. The replacement of lines with text may be confusing, especially when there are many lines.


Line graph of stocks with biggest gains on September 23, 2015. The stock information is written inside the lines starting at the time of day of the reporting.
Line graph of stocks with biggest gains on 9/23/15. The stock information is written inside the lines starting at the time of day of the reporting.

When it comes to sets and categories, Brath employs Venn Diagrams to display multiple categories of data. His Venn Diagram of the 114th US Senate clearly shows categories of gender, number of terms, age, education, ethnicity, as well as the party of each senator using a mix of Venn Diagram divisions and text variables.

In the figure below, you can tell that Mazie Hirono is a female (purple), Democrat (left leaning text), over age 65 (ALL CAPS), first term senator (not bold) with an advanced degree (underline), and an Asian American heritage (serif). This visualization gives you both overall trends and specific information about individuals.


Venn Diagram of the 114th U.S. Senate
Venn Diagram of the 114th U.S. Senate

In order to be readable, Venn Diagrams should be limited to a few variables with a minimal number of categories. Brath describes options for encoding larger numbers of variables and describes typographic graph types, including text-based graph-based layouts, scatterplots, mosaic plots, and stacked bar charts. He even discusses using different visual encodings for each character in a label or word.


Visualization with Text

While the bulk of the book deals with text and data visualization, Brath demonstrates a number of innovations for applying visualization techniques in text. He incorporates underlining, bolding, or highlighting long text labels to highlight the proportions or counts associated with list-based text data.


Augmenting text with visual elements. The underlines show the article length (in 1,000s of words), while the length of the highlighting shows the number of page views (in 1,000s of views).
Augmenting text with visual elements. The underlines show the article length (in 1,000s of words), while the length of the highlighting shows the number of page views (in 1,000s of views).

When it comes to prose, Brath offers techniques to support skimming for “key ideas and content overview.” An excellent example is an excerpt from Jane Austen’s Emma, where key words are highlighted with bold text. Text readability decreases with increasing complexity of the symbols, but Brath notes that digital systems can toggle between the original text and the enhanced text.


Excerpt from Jane Austen’s “Emma” with key words highlighted by bold text. This facilitates text skimming while retaining the important content.
Excerpt from Jane Austen’s “Emma” with key words highlighted by bold text. This facilitates text skimming while retaining the important content.

Brath shows a creative method for displaying text to assist pronunciation, while retaining the original spelling. The original letters that assist pronunciation are shown in black; vowels are underlined if they are pronounced with a hard sound; original letters that do not assist pronunciation are shown in gray; with replacements for the letters are shown in red. This is another technique where toggling between the original text and the enhanced text improves the experience.


Text visualization designed to enhance pronunciation.
Text visualization designed to enhance pronunciation.

Despite his enthusiasm for the use of text, Brath is transparent and honest about the challenges. Using text in data visualizations can take up precious graphical real estate. The use of too many text variables to encode multiple attributes can make text difficult to read. And language presents its own challenges. Languages may not use the same characters, order letters in the same way, or read in the same direction. Thus, text lacks some of the universality of symbols.

Brath’s work incorporates an expansive view of text visualization. He shows how the thoughtful integration of text can increase the designer’s options. His examples are wide ranging and innovative. Visualizing with Text will change the way you think about text, the use of text in visualizations, and the application of visual enhancements to text. Brath has thought deeply and long about visualizing text and presents a framework that integrates the visual variables, text variables, text non-type visual attributes, scope, and text layout. His comprehensive exploration and diverse examples stir the imagination.

All images in this book review were provided courtesy of Richard Brath, who released them under the CC BY-SA 4.0 license. These and other images can be found at … https://richardbrath.files.wordpress.com/2020/10/diagrams_and_figures_from_visualizing_with_type_richard_brath_cc_by_sa_4.0_20201023e.pdf


Resources


91 views0 comments

434-414-1085

401 Water St. E, Charlottesville VA 22902

  • Twitter
  • Facebook

©2021 by Center for Civic Innovation. Proudly created with Wix.com

bottom of page