Content uploaded by Garreth W. Tigwell
Author content
All content in this area was uploaded by Garreth W. Tigwell on Sep 20, 2023
Content may be subject to copyright.
Data Visualization Accessibility for Blind
and Low Vision Audiences
Chloe Keilers, Garreth W. Tigwell, and Roshan L. Peiris(B
)
School of Information, Rochester Institute of Technology, Rochester, NY 14623, USA
{cmk4967,garreth.W.Tigwell,roshan.peiris}@rit.edu
Abstract. While data visualizations have the potential to convey vast
quantities of information, they are not always accessible to audiences
with vision impairments. We prepared and distributed an online survey
to blind and low vision adults to investigate the accessibility of data visu-
alizations across the following five mediums—computers, phones, tablets,
paper, TVs. After analyzing 45 survey responses, we identified that the
inaccessibility is pervasive and that people want to interpret data inde-
pendently. At present, data visualizations are largely inaccessible to blind
and low vision users; however, it is possible to improve accessibility with
intentional design.
Keywords: BLV blind low vision graphics visualization
accessibility
· · · · ·
1 Introduction
Data visualizations are ubiquitous, from our paper’s bar charts to the latest
health department pandemic statistics [11,26]. Visualizations use visual features
to convey meaningful information to an audience [20,29]. People who are blind,
have low vision, color vision deficiency, etc. might not be able to access or per-
ceive visualization features, and, therefore, miss what the visualization conveys,
which could be a critical communication barrier. An inaccessible pandemic graph
released by a newspaper or health outlet may deprive blind and low vision peo-
ple of vital information in the middle of a health crisis [11]. Designing accessible
visualizations and tools not only make this information available to a wider audi-
ence but may even be required to meet legal and Americans with Disabilities
Act (ADA) requirements [1]. For visualizations to be accessible, they must first
be detectable and then provide summaries of data trends, freedom to explore
details, and meaningful context [17].
Some options for blind people to access graphs include screen readers, tactile
charts, and magnification. Screen reader users could access data visualization via
a high level summary in alternative text (alt text) [16] or by navigating tabular
data [25]. Accessing charts with screen readers takes twice as long for users to
understand when compared to non-screen reader users [24]. Tactile devices con-
vey charts through embossment, haptics, and braille annotations [14]; however,
Official version can be accessed at the following link: https://doi.org/10.1007/978-3-031-35681-0_26
these are expensive, difficult to find, and time-consuming to create or auto-
mate [9].
In this research, we investigate the current state of data visualization acces-
sibility across several mediums from a Blind and Low Vision (BLV) users’ per-
spectives and propose several recommendations for approaching accessible data
visualization. For this purpose, we created and distributed a survey to answer
the research question “What are the accessibility perspectives of a BLV users
on access data visualizations across different mediums?”. As our main contribu-
tions, in our findings with 74 BLV survey participants, we found that blind and
low vision people consider data visualizations across all mediums to be inacces-
sible and poorly designed and, these users desire more independent alternatives
to access data visualizations.
2 Related Work
Data visualization is the visual representation of a data set [29]. Data visual-
izations appear everywhere from daily newspapers to online scientific papers,
ranging from simple line graphs to complex interactive graphs with multiple
axes. There are many approaches to make data visualization accessible for blind
or low vision audiences, both technological and methodological. While there are
many studies that investigate the specific technologies and methods used to
access a data visualization, there are few qualitative studies that investigate the
current practices and perspectives of blind and low vision people.
2.1 Data Visualization
Many data visualizations appear in professional settings such as scientific liter-
ature, business reports, and engineering diagrams, and those visualizations are
powerful tools that have become more available with the rise and increasing
prevalence of computer graphics and mass media [10]. Data visualizations are
used to inform and enhance what the data is saying, from a phone’s battery
status icon to a timeline graph of stocks to the seven day rolling average of
COVID-19 cases [11,26]. Data visualizations have informed populations, high-
lighted issues, and help changed health care throughout history [7,23]. Two
early examples include John Snow’s map of 1854 cholera outbreaks that showed
a link between the disease and water [10] and Florence Nightingale’s 1858 pie
chart showing preventable deaths as a result of lack of sanitation as the leading
cause of army fatalities [10]. Outside of healthcare, data visualizations help peo-
ple make better informed decision in everyday life, such as the New York City
metro map transporting commuters and tourists alike efficiently [21].
Ware [29] describes comprehending data visualization first by having a “prob-
lem to be solved,” then carrying out a “visual search” to answer the initial query.
A graph exists to display data for the users to draw their own conclusions. The
user’s comprehension and conclusion rely heavily on the visual characteristics
of the graphs [20]. To make data visualizations more accessible to blind people,
the format and visual characteristics of the graph must be available, while also
providing blind people the freedom to explore the graph to answer their own
questions.
2.2 Current Technology for Accessibility
Numerous methods and technologies aim to make data visualizations accessible
to BLV people by using either tactile or aural interfaces. Kim et al. [17] surveyed
available literature and put forward a visualization accessibility model where
accessible is “notifying the existence of a chart, providing an overview of the
chart, offering details only when requested, [and] conveying the context when
necessary and helpful.” Several methods use combinations of these technolo-
gies. When viewed under the accessibility model, each technology has different
strengths and weaknesses.
Screen Readers: Alt Text: Screen readers are widely available as they need no
particular hardware beyond a computer and speaker. The most common method
used by screen readers to handle images is to read the image’s alt text. Jung
et al. [16] surveyed and interviewed blind people on how alt text is currently
used and could be improved for data visualizations by testing four styles of alt
text: brief descriptions, detailed descriptions, data trends, and raw data points.
Jung et al. concluded that more detailed alt text and access to raw data points
were the preferred options. Participants mentioned their desire to access data
freely and to know the chart type and visual features [16]. A single alt text
may not provide sufficient freedom to explore a chart on its own; however, it
could provide a high-level summary while being leveraged with other methods
to explore the data.
Screen Readers: Others Screen readers in combination with other tools
increase opportunities to read graphs. A study by Godfrey et al. [12] explores
the creation and exploration of data visualization by screen reader users in the
R statistical programming language. Using R and screen readers, participants
could not only create graphs but also explore the data interactively. While their
study showed screen readers’ important utility, they are not perfect. Sharif et al.
[24] found that if screen readers detect a graph it takes twice as long for users to
understand, and with less accuracy, when compared to non-screen reader users.
This is due to screen readers, relatively one-dimensional devices, reporting every
label and x-y coordinate for each data point. Additionally, the performance and
capabilities of screen readers can have an impact on the user’s understanding.
Beal et al. [2] studied an iPad math app and noted that VoiceOver screen readers
read fraction “1/4” as “one slash four” as opposed to “one over four” or “one
fourth”. Screen readers do not read or speak the same as humans do. While
screen readers are widely available, they alone cannot provide full access to data
visualizations.
Tactile Devices. Tactile devices convey graphs through textures, embossments,
braille annotations, and 3-D prints [14]. Unfortunately, tactile charts are not only
expensive and more difficult to find, but they are also more time-consuming to
create as Engel and Weber [9] show when attempting to automate the task.
Touch screens are also popular due to their widespread commercial availabil-
ity. They open more visualizations opportunities, especially when combined with
audio or haptic methods. Beal and Rosenblum [2] tested an iPad app with tactile
graphs to teach blind and low vision students mathematics, taking advantage of
Apple’s VoiceOver to help students interact with the display. In this case, the
tablet functions as a screen reader with a touch interface to provide more details
about graphs.
Low Vision Aids. Low vision aids include various magnification devices which
may improve accessibility for low vision users [18]. Additionally, most computers
and phones have built-in zoom options or accessibility features to change the
colors of a screen. Low vision and color blind users use magnifiers to enlarge or
enhance the image.
Optical Recognition. Optical character recognition (OCR) pulls text from
pictures to make it accessible to computers and by extension to people with
screen readers. The U.S. Government Printing Office found that, for images
with printed text, the OCR accuracy rate hovers above 98% [3]. For more com-
plicated images and environments, artificial intelligence (AI) has been applied
to improving image processing and recognition. Granquist et al. [13] tested two
AI vision aids to read warped text, such as a soup can label, from images with
13-57% accuracy.
2.3 Creation of Data Visualization
When creating data visualizations, there are two ways to make them accessible:
first, give the power and tools to blind and low vision audience to create data
visualizations themselves from raw data, second, educate and support creators to
design more accessible visualizations. In addition to Godfrey et al.’s [12] R sta-
tistical programming language, there are several tools that enable screen readers
to interpret code that create and read data visualizations [5,27].
Heer et al. [15] discussed a wide array of different data visualization types
and what considerations should take place when choosing how to visualize and
present data. They strongly encourage simpler data visualizations when possible
and emphasized that “the DNA underlying all visualizations remains the same;
the principled mapping of data variables to visual features such as position, size,
shape, and color” [15]. While complex and interactive data visualizations are
harder to make accessible, this “mapping” of data to visual features is also an
important consideration to making any visualization accessible. [6] recommended
some design guidelines based on how blind “Orientation & Mobility” instructors
build a mental model of space.
After a data visualization is created, a set of heuristics principles could be
applied to check and fix inaccessible designs. Web Content Accessibility Guide-
lines (WCAG) 2.0 have been adopted by U.S. courts as the website accessi-
bility standards [28]. The first WCAG originated when the internet was less
media heavy and has been slow to adapt to changing technology and internet
usage [19]. More recently, Elavsky et al. [8] created a set of heuristic princi-
ples for visualization accessibility called Chartability. Chartability starts with
four heuristics principles from WCAG: Perceivable, Operable, Understandable,
and Robust (POUR); while introducing three more principles: Compromising,
Assistive, and Flexible (CAF).
2.4 Summary
As observed here, accessible data visualizations is a topic that has been explored
widely. In summary, while many researches have explored the usability of exist-
ing methods for BLV people accessing data visualizations, several studies have
proposed methods that use techniques such as haptics for presenting accessi-
ble data visualizations to BLV people. Thus, in our research, we examine BLV
peoples’ perspectives on the accessibility of data visualizations across several
mediums and discuss their preferences and requirements, as well as present rec-
ommendations for when developing new methods in the future.
3 Survey
To determine the most popular, accessible, and user-friendly method to access
graphs, we created and distributed a survey for adult blind and low vision par-
ticipants. Our survey gathered information on current technology usage and
data visualization practices. We had 25 questions—five multiple-choice and 20
text. We scrubbed text responses and grouped shorter quantitative answers into
appropriate categories. Longer text fields required more qualitative thematic
analysis [4]. We posted and emailed the survey to various internet groups and
organizations for blind and low vision people. The survey was anonymous, but
the respondent could enter a $15 USD lottery if they chose to.
We collected 74 responses over two months. Out of the 74, 45 valid responses
were identified where, the rest were removed due to being incomplete, repeated or
was identified as spam. All participants were adults, 18 and over, with a median
age 45-54 and a mode age 65 or older. The gender split was 56% male and 44%
female. There were 14 totally blind participants, 19 legally blind participants.
Of the rest, 5 identified as totally/legally blind and 7 as either low vision. At
least 19 participants have been blind since early childhood. Educational back-
grounds ranged from high school to doctoral degrees; three-quarters had some
post-secondary education. Half (24) reported having no science, technology, engi-
neering, or math (STEM) experiences; ten currently work or study STEM; eight
previously worked or studied STEM; and three were unsure if they had any
STEM experiences. All but two participants use smartphones. One-third (14) of
participants reported using tablets. Over 85% (39) of participants use comput-
ers. Two-thirds (30) of participants used screen readers, including JAWS (25),
VoiceOver(9), NVDA(8), and others. Thirteen participants used more than one
screen reader, i.e. JAWS on computers and VoiceOver on iPhones.
4 Findings
Fig. 1. Frequency of graph encounters on different mediums by number of responses.
We asked participants about their usage of the following five mediums—
computers, phones, paper, tablets, and TVs. Participants rated how often they
encounter data visualizations on each medium (Fig. 1); computers and phones
have the most frequent encounters. Participants rated how often graphs are
accessible and inaccessible on a scale of “1 Never” to “5 Always”. We calcu-
lated net rating by comparing accessibility “Always” responses to inaccessibility
“Never” responses (Fig. 2). Based on net rating, all mediums appear largely inac-
cessible; however, computers and phones might be considered the most accessi-
ble. Methods used by each type of medium include:
– Computers were the most popular with 31 answers. Screen readers via alt
text, optical character recognition (OCR), or third-party extension such as
Fig. 2. Average accessibility where on a scale, positive 5 indicates always accessible
and negative (5) indicates never accessible. The average figure is shown as the net
rating where within brackets indicate a negative value
SAS Accelerator are the primary method used to access visualizations on com-
puters. People without screen readers may use large monitors that adjust mag-
nifications. Other responses mentioned braille display, zoom software, sighted
help, and spreadsheet data.
– Phones had 24 responses. Ten used screen readers using either alt text, OCR,
or image description. Other responses mentioned sighted help, audio chart
descriptions, and sonic graphs in a weather app. People without screen readers
do not consider phones as accessible as people with screen readers due to
smaller screens.
– Paper had nine responses: three used braille or tactile graphs; four asked for
sighted help; two scanned graphs. One of the tactile responses described an
old device from the ’70s that traced the shape of a graph. For people with
low vision who use a magnifying tool, paper also reduces eye fatigue.
– Tablets were the second least popular medium with only five responses. Four
used a screen reader, and one used an audio chart description.
– TVs were the least popular medium, with only four responses. Each person
used something different: audio description, Roku, Google Chromecast, or
Amazon Firestick.
4.1 Importance of Data Visualization Accessibility
Over half of the participants said understanding data visualizations is impor-
tant particularly for work scenarios, financial interests, and medical informa-
tion. Forty-one participants answered the question “How important is accessing
and understanding a graph to you?”. This question had no specific instructions
on how to answer; however, when assessing how important accessing the data
visualization were to the participants, most of the answers were some variations
of “Very Important,” “Rarely,” or “It depends.” These answers were converted
into a scale system as shown in Fig. 3. Twenty, nearly half, of participants said
understanding graphs is “Very important” and six said “Somewhat or mod-
erately important.” Five participants each answered either “It depends on...,”
“Low importance,” or “Not important.” Since it was an open-ended question,
some participants expanded on their answers, details of which were added to the
analysis.
Fig. 3. Responses of how important accessing and understanding is a graph from
“Never” to “Very Important.”
Thirty-six responses provided examples of when accessing data visualization
is most important. Nine participants mentioned on the job scenarios; six men-
tioned understanding financial, investment, and banking information; and others
mentioned accessing medical information, temporal data, and mass communica-
tion. Participants really care about the data visualization when the context is
important to them and their personal interests. Some comments emphasized the
importance of context and access to information: “Very important as these are
often where the core information of reports are presented.” (P02) where as P23
said, “If I need to obtain information from said graph to answer questions.” Fur-
thermore, P24 added “When there are not other ways to get the data or if trying
to understand trends.”
Participants are invested when the context is important to their personal
interests. P44 said, “It is very important to read chart when data is not available
in text format. If I cannot get data, chart remains only available option to read
data.” Another (P37) said, “When I deem the information important, such as
COVID stats, etc.” A participant (P05) concluded, graphs are “very important
as these are often where the core information of reports are presented.”
The desire for more accessible data is related to the desire for inclusion and
independence. P08 called for more open and accessible data, saying “Democra-
tization of knowledge and finding ways to be more inclusive through quantitative
or qualitative research.” Another (P06) emphasized how useful visualizations are
as a “visual shorthand” and said, “A graph exists to convey information...A good
presentation should be able to convey all the information in the graph in words
and text with minimal loss of emotional impact.”
4.2 When a Data Visualization Is Inaccessible
When a data visualization is not easily accessible, many participants attempt
more than one technique. One participant, P25, listed their options: “Ignore it,
look for a text description, get someone to summarize, or emboss it in Braille.”
Twenty-two participants said they would ask for sighted help, though 16 said
this is not their first choice. P32 said, “I ask somebody to describe it or try to
understand the content without it.” A few mentioned advocating for themselves
by contacting the creator for either an alternate format or access to the raw
data. This takes additional time and may not provide enough information to
interpret graphs. As one participant, P05 said, “Seek information in another
format e.g. the raw data from which the chart was created. However, the charts
often interpret the raw data so getting the raw data does not highlight the key
message that the chart is presenting.” Eleven participants said they would skip
or ignore the visualization, but this is their last resort. Here, P26 said, “If I can,
I skip over it.”
4.3 Suggested Areas of Improvements
Participants are frustrated by poor designs that hinder their access to data
visualization. As such, there were 36 responses and of those only five said “None”
or “Don’t know.” The most common suggested responses were:
– Include better screen reader support across all software and mediums. “I
would like screen readers to have support for it” (P23)
– Create narrative summaries, either in alt text or context of the article, “Better
descriptions of graphs or charts. Alt text would also be useful” (P26).
– Provide an audio format, either descriptions or sonic graphs, “Better alt text
or descriptions that can be interacted with to pay better attention to numbers
choice of audio representation” (P45).
– Make data navigation easier, “Wouldlikemoreindependent waytoaccess
them without sighted assistance ” (P31).
– Design with accessibility and inclusiveness in mind, i.e. choose high contrasts
colors and textures; use large fonts; label axes, rows, and columns “Various,
especially thin lines, are usually difficult to discern. Maybe alternative methods
to distinguish lines on a graph would help” (P24).
– Enhance tactile graphics, “...Make tactile displays which can show graphs
really affordable (and refreshable)” (P29).
– Provide raw data, Excel or spreadsheet version, “Always build in excel so that
data can be navigated easily...” (P29).
4.4 Other Comments
At the end of the survey there was a text box for participants to share closing
thoughts and comments about the subject. One participant emphasized how
useful graphs are as a “visual shorthand” and said that a good presentation or
article should equally convey all the information in the text itself.
“A graph exists to convey information, not emotion. It’s not art that needs
to be seen to be appreciated. It’s just visual shorthand, a model to convey
information in an easy way. A good presentation should be able to convey
all the information in the graph in words and text with a minimal loss of
emotional impact.” (P06)
A few other participants expressed the desire to read graphs without help
and their frustration that creators do not provide accessible format for the data.
“It would be great to be able to do them without help.” (P19)
“Most graphic and charts can be created in Excel and then provided in an
accessible form, but most sighted folks do not do this.” (P31)
5 Discussion
We found BLV people use many methods and technologies to access data visu-
alizations. We gauged which of the five mediums (computers, phones, paper,
tablets, TVs) are most accessible. Types of methods and technologies vary
according to the user’s level of vision, access to technology, and available medi-
ums. These methods may be divided between screen reader users (SRUs), who
are predominantly legally or totally blind, and non-SRUs, who generally have
low vision.
SRUs use third party screen reader extensions, audio descriptions, and occa-
sionally sighted help to access data visualizations. They do not consider data
visualizations on any medium very accessible. Non-SRUs generally use various
magnification approaches and sighted help as well. They perceive all mediums to
be slightly more accessible. TVs and tablets are not popular among either group,
perhaps due to few users or due to visualizations on these mediums being espe-
cially challenging. Five participants mentioned using tactile and braille methods
in addition to screen readers.
Multiple alternatives to provide access are important because there is no
one universal method that works for everyone. Redundant non-visual measures
to access graphs include: alt text descriptions, figure captions, tabular forms
of data, and possibly tactile, haptic, or sonic forms of graphs. Sharif, et al.,
recommends auto-generated alt text via artificial intelligence [24]. When auto-
generated text is accurate, this both relieves pressure on the graph’s creator and
also enables independent exploration for users. Jung suggested creating hidden
HTML data tables and mentioning their location in the graph’s alt text [16].
This does not work in all circumstances, especially when there is too much data.
Not every user appreciates having access to tabular data.
5.1 Pervasive Inaccessibility
Inaccessibility is pervasive; data visualizations are not free from their environ-
ments. Users who are blind or have low vision must navigate past challenging
obstacles to recognize there is a graph in the first place. There are physical, men-
tal, and emotional tolls from inaccessibility. Participants who encounter graphs
spend both time and effort trying to understand a graph and advocating for
better support. When the graph is not important, it is easier to skip.
5.2 Independent Interpretation
Everyone has multiple methods for accessing data visualizations, but many
emphasize the desire to interpret data independently. Half of the participants
ask for sighted help; however, for three quarters of those it is a final resort, only
to be used when the data is important. Sighted help is also not always reliable.
Those who have experience with braille charts also prefer them for the power and
independence they provide. Frequently, participants express the desire for easier
navigation and the power to manipulate the graph in order to access information.
Tactile charts provide independent access to visualizations; however, they are
difficult to find for adult users. Although 3-D printing is becoming more acces-
sible to commercial markets, it is still rare [14]. Haptic devices are becoming
more widespread in modern smartphones and could apply to a variety of visu-
alizations [22]. Current tactile graphs and haptic prototypes require investment
to buy equipment, and neither technology is as prevalent as screen readers.
5.3 Design Frustration
Poor visualization design hinders access. Participants feel this could be avoided
with more mindful planning. Most frustrations with design could be avoided if
graph creators check their visualizations using heuristic guidelines. Web Con-
tent Accessibility Guidelines have been adopted by U.S. courts as the website
accessibility standards [28] and might be a good place to start for guidance.
More recently, Elavsky, et al., created a set of heuristic principles for visualiza-
tion accessibility called Chartability [8], which can help creators avoid and fix
inaccessible design.
6 Recommendations
Creators should use heuristics to improve data visualizations and provide alter-
nate ways of accessing data such as ensuring adequate text descriptions, provid-
ing access to data tables, and investigating more versatile tactile means.
6.1 Heuristics Checklist
Poor visualization design hampers access. Most design obstacles could be avoided
if creators intentionally design and check their visualizations using heuristic
guidelines. Existing and future heuristics should be verified with end users to
ensure results meet expectations.
Taking an heuristic approach might start with Web Content Accessibility
Guidelines (WCAG) 2.0. These were adopted by the U.S. courts as website acces-
sibility standards [28]. More recently, Elavsky et al. created a set of heuristic prin-
ciples for visualization accessibility called Chartability [8]. Chartability expands
on WCAG 2.0’s four heuristics principles–Perceivable, Operable, Understand-
able, and Robust–while introducing three additional principles–Compromising,
Assistive, and Flexible.
6.2 Alternate Access
Data visualizations are most commonly shared through images or interactive
codes; however, there are redundant non-visual measures to access graphs,
including alt text descriptions, caption descriptions, tabular forms of data, and
possibly tactile, haptic, or sonic forms of graphs. Jung et al. recommends com-
prehensive alt text measures [16].
Sharif et al. recommends auto-generated alt text via artificial intelligence
[24]. When auto-generated text is accurate, this both relieves pressure on the
graph’s creator and also increases independent exploration for users. Jung et al.
suggested creating hidden HTML data tables and mentioning their location in
the graph’s alt text [16]. This does not work in all circumstances, especially when
there is too much data. Not every users appreciates having access to tabular data.
Tactile charts provide independent access to visualizations; however, they are
difficult to find for adult users. Although 3D printing is becoming more accessible
to commercial markets, it is still rare [14]. Haptic devices are becoming more
widespread in modern smartphones and could apply to a variety of visualizations
[22]. Current tactile graphs and haptic prototypes require investment to buy
equipment and neither technology is as prevalent as screen readers.
7 Conclusion
Blind and low vision users consider the current state of data visualization to be
unsatisfactory-most users consider visualizations easier to skip. Methodologies
and technologies vary according to the user’s level of vision, access to technol-
ogy, and available mediums. Blind participants typically rely on screen readers,
but show an interest in tactile formats. Low vision participants use magnifica-
tion tools, but cannot control how well graphs are designed. Both groups will
skip visualizations, ask sighted people for help, or contact graph creators for
details when initial methods fail. Graph creators should be more intentional.
By researching heuristics, alternatives, and new modes of accessing graphs, it is
possible to make data visualizations more accessible.
References
1. Administration, G.S.: Section 508 of the Rehabilitation Act (1973). https://www.
section508.gov/
2. Beal, C.R., Rosenblum, L.P.: Use of an accessible iPad app and supplemental
graphics to build mathematics skills: feasibility study results. J. Vis. Impairment
Blindness (Online) 109(5), 383 (2015)
3. Booth, J.M., Gelb, J.: Optimizing OCR accuracy on older documents: a study of
scan mode, file enhancement, and software products (2006)
4. Braun, V., Clarke, V., Hayfield, N., Terry, G.: Thematic analysis. In: Liamputtong,
P. (ed.) Handbook of Research Methods in Health Social Sciences, pp. 843–860.
Springer, Singapore (2019). https://doi.org/10.1007/978-981-10-5251- 4 103
5. Center, S.H.: Introduction to the SAS graphics accelerator. https://documentation.
sas.com/doc/en/gracclug/1.0/p04o83muel10yen12ptmogjpapez.htm (2021)
6. Chundury, P., Patnaik, B., Reyazuddin, Y., Tang, C., Lazar, J., Elmqvist, N.:
Towards understanding sensory substitution for accessible visualization: an inter-
view study. IEEE Trans. Visual Comput. Graphics 28(1), 1084–1094 (2022).
https://doi.org/10.1109/TVCG.2021.3114829
7. Crisan, A.: The importance of data visualization in combating a pandemic. Am.
J. Public Health 112(6), 893–895 (2022)
8. Elavsky, F., Bennett, C., Moritz, D.: How accessible is my visualization? evaluating
visualization accessibility with chartability. Eurograph. Conf. Visualiz. (EuroVis)
2022 41(3), 14522 (2022)
9. Engel, C., Weber, G.: Improve the accessibility of tactile charts. In: Bernhaupt, R.,
Dalvi, G., Joshi, A., K. Balkrishan, D., O’Neill, J., Winckler, M. (eds.) Human-
Computer Interaction - INTERACT 2017, vol. 10513, pp. 187–195. Springer Inter-
national Publishing, Cham (2017). https://doi.org/10.1007/978-3-319- 67744-6 12
10. Friendly, M.: A brief history of data visualization. In: Handbook of Data Visualiza-
tion, pp. 15–56. Springer, Heidelberg (2008). https://doi.org/10.1007/978-3-540-
33037-0 2
11. Gleason, C., et al.: Disability and the Covid-19 pandemic: using twitter to under-
stand accessibility during rapid societal transition. In: The 22nd International
ACM SIGACCESS Conference on Computers and Accessibility, pp. 1–14. ACM,
Virtual Event Greece (2020). https://doi.org/10.1145/3373625.3417023
12. Godfrey, A.J.R., Murrell, P., Sorge, V.: An accessible interaction model for data
visualisation in statistics. In: Miesenberger, K., Kouroupetroglou, G. (eds.) Com-
puters Helping People with Special Needs, vol. 10896, pp. 590–597. Springer, Cham
(2018). https://doi.org/10.1007/978-3-319- 94277-3 92
13. Granquist, C., Sun, S.Y., Montezuma, S.R., Tran, T.M., Gage, R., Legge, G.E.:
Evaluation and comparison of artificial intelligence vision aids: Orcam MyEye 1
and seeing AI. J. Vis. Impairment Blindness 115(4), 277–285 (2021)
14. He, L., Wan, Z., Findlater, L., Froehlich, J.E.: Tactile: a preliminary toolchain for
creating accessible graphics with 3D-printed overlays and auditory annotations. In:
Proceedings of the 19th International ACM SIGACCESS Conference on Computers
and Accessibility, pp. 397–398. ACM, Baltimore Maryland USA (2017). https://
doi.org/10.1145/3132525.3134818
15. Heer, J., Bostock, M., Ogievetsky, V.: A tour through the visualization zoo. Com-
mun. ACM 53(6), 59–67 (2010). https://doi.org/10.1145/1743546.1743567
16. Jung, C., Mehta, S., Kulkarni, A., Zhao, Y., Kim, Y.S.: Communicating visual-
izations without visuals: investigation of visualization alternative text for people
with visual impairments. IEEE Trans. Visual Comput. Graphics 28(1), 1095–1105
(2022). https://doi.org/10.1109/TVCG.2021.3114846
17. Kim, N.W., Joyner, S.C., Riegelhuth, A., Kim, Y.: Accessible visualization: design
space, opportunities, and challenges. Comput. Graph. Forum 40(3), 173–188
(2021). https://doi.org/10.1111/cgf.14298
18. Latham, K.: Benefits of low vision aids to reading accessibility. Vision. Res. 153,
47–52 (2018). https://doi.org/10.1016/j.visres.2018.09.009
19. Lazar, J., Dudley-Sponaugle, A., Greenidge, K.D.: Improving web accessibility: a
study of webmaster perceptions. Comput. Hum. Behav. 20(2), 269–288 (2004).
https://doi.org/10.1016/j.chb.2003.10.018
20. Lee, S., Kim, S.H., Hung, Y.H., Lam, H., Kang, Y.A., Yi, J.S.: How do people
make sense of unfamiliar visualizations?: a grounded model of novice’s information
visualization sensemaking. IEEE Trans. Visual Comput. Graphics 22(1), 499–508
(2016). https://doi.org/10.1109/TVCG.2015.2467195
21. Lloyd, P.B., Rodgers, P., Roberts, M.J.: Metro map colour-coding: effect on usabil-
ity in route tracing. In: Chapman, P., Stapleton, G., Moktefi, A., Perez-Kriz, S.,
Bellucci, F. (eds.) Diagrams 2018. LNCS (LNAI), vol. 10871, pp. 411–428. Springer,
Cham (2018). https://doi.org/10.1007/978-3-319- 91376-6 38
22. Paneels, S., Roberts, J.C.: Review of designs for haptic data visualization. IEEE
Trans. Haptics 3(2), 119–137 (2010). https://doi.org/10.1109/TOH.2009.44
23. Segel, E., Heer, J.: Narrative visualization: telling stories with data. IEEE
Trans. Visual Comput. Graphics 16(6), 1139–1148 (2010). https://doi.org/10.
1109/TVCG.2010.179
24. Sharif, A., Chintalapati, S.S., Wobbrock, J.O., Reinecke, K.: Understanding screen-
reader users’ experiences with online data visualizations. In: The 23rd International
ACM SIGACCESS Conference on Computers and Accessibility, pp. 1–16. ACM,
Virtual Event USA (2021). https://doi.org/10.1145/3441852.3471202
25. Torres, M.J.R., Barwaldt, R.: Approaches for diagrams accessibility for blind peo-
ple: a systematic review. In: 2019 IEEE Frontiers in Education Conference (FIE),
pp. 1–7. IEEE, Covington, KY, USA (2019). https://doi.org/10.1109/FIE43999.
2019.9028522
26. Valencia, S., Kirabo, L.: Twitter, Covid-19, and disability: what worked and what
didn’t. XRDS 28(2), 20–23 (2022). https://doi.org/10.1145/3495255
27. Yu, W., Kangas, K., Brewster, S.: Web-based haptic applications for blind people
to create virtual graphs. In: 11th Symposium on Haptic Interfaces for Virtual
Environment and Teleoperator Systems, 2003. HAPTICS 2003. Proceedings, pp.
318–325. IEEE Comput. Soc, Los Angeles, CA, USA (2003). https://doi.org/10.
1109/HAPTIC.2003.1191301
28. Walker, K.: Modern day technology: not accessible to all, but necessary to navigate
this society. Syracuse J. Sci. Tech. L. 35, 98 (2018)
29. Ware, C.: Information visualization: perception for design. MA, third edn, Inter-
active technologies, Morgan Kaufmann, Waltham (2013)
