When asked what they believed were the one to three most important issues facing data visualization [Figure 18], respondents indicated the lack of data visualization literacy, the lack of awareness of the impact of data visualization, and data visualization not having a seat at the table as key issues. These top three issues were the same for both new and seasoned data visualizers.
Individuals who identified as being part of a historically underrepresented or disadvantaged racial or ethnic group also indicated lack of data visualization literacy and awareness of its impact as their top two issues, and included diversity in tech as their third most important issue (up from sixth in the overall analysis).
Speaking to the gaps in data visualization literacy and awareness of impact, nearly half of the survey respondents indicated there was something about the data visualization work they do that people they work with don’t understand. The primary theme, identified by nearly one in three respondents, was that others don’t realize how much time and effort it costs to visualize data well. Numerous respondents expressed this in few words: “Quality takes time.” Others elaborated on their process, indicating “They don’t understand that I don’t just “pull the report”… it’s data from like 20 different locations, that needs to be standardized, formatted, cleaned, then presented.”
Another primary concern of approximately one in five respondents is the lack of data savvy. Understandably, this category overlaps with the time investment of data visualization, but also encompasses understanding and processing sometimes unfamiliar data. People don’t always know “That understanding the underlying data on which the viz is built can be complex and [it’s] not easy to accurately portray in a software platform.” Others included comments about the value of data processing and analysis, indicating “how statistical properties of data (should) inform visualisation choices is not always appreciated.”
A substantial challenge respondents face is explaining that data visualization design is a key part of the process. “They think it’s just about being ‘pretty’ and don’t understand that every element has a purpose and was carefully thought through.” In addition, there seems to be limited understanding that for visualizations to be effective, they need to be tailored for the audience and the situation. “Understanding the users’ context and their goals/needs before just choosing a visualization method.”
At the intersection of design and value were quotes such as: “Carefully crafted visuals can greatly reduce the time to insight. Design plays a huge part in whether or not the end user can make sense of data.” emphasizing the multifaceted uses of data visualization. Respondents noted that others don’t always see value in the capacity for data visualization to communicate important information clearly and gain insights for decision-making and problem solving.
Finally, respondents identified challenges with change management, their colleagues’ and leaders’ “reluctance to try new methods of visualizing data.” They mentioned allegiance to certain software and visualization types, including bar charts, pie charts, line graphs, and tables.
What else should we know about the state of the data visualization industry?
DVS 2021 SOTI survey respondents recognize there are broad issues impacting the data visualization community. Our top three issues were lack of data visualization literacy, the lack of awareness of the impact of data visualization, and data visualization not being involved in early conversations about data collection, structure, analysis, and reporting. We also identified and put words to the struggles with what others just don’t get about the data visualization work we do, trying to find new ways to tell leaders that quality data visualization takes time. Dashboards and tables are not always the answer, and complex data take careful consideration in how they should be presented. This time spent understanding the data and considering design elements can be critical for presenting clear information to the end user and, ideally, providing them with data they can readily use to drive decision-making. Data visualizers want to meet the needs of clients, create impactful visualizations, and change the world but often feel stymied, potentially presenting challenges for growth of the industry.
A total of 66 questions were designed to elicit responses across topics including demographics, roles and tasks, experience and compensation, tools and charts, challenges and changes, COVID-19 impacts, looking toward the future, and role-specific questions for freelancers/consultants/independent contractors, individuals who work in organizations, academics, students, and hobbyists. All questions were optional with the exception of the first asking about role, which allowed for the survey to direct participants to selected items related to identified role(s); individuals could still select “Prefer not to answer” for this required question to move forward with the remainder of the survey. Individuals who reported holding multiple roles were asked questions from each relevant section; the most questions asked to an individual participant with many roles was 60. On average, participants were asked 47 questions (mean = median), answered 43 questions (median; mean = 39) and completed the survey in 11 minutes (median).
Key Survey Items
A total of 2,165 respondents provided usable data for analysis in the survey and 1,788 reached the end of the survey. As described above, this report does not detail all responses to each item, but rather seeks to answer a few key questions about the state of the industry. Here we outline the survey items that will answer each question.
Who are we?
To provide context to our global network, visualizers answered a series of demographic questions, including DVS membership status. It should be noted that the membership structure of DVS changed in June 2021; individuals had to opt into membership and accordingly, distribution of members and non-members taking the survey has shifted this year. Visualizers also shared their location. We plotted the city and country locations on a map, with each location segmented into buckets to approximate how many visualizers are in a location; the buckets align with buckets created for the 2020 map images. We also have global representation in terms of languages and asked visualizers to share the languages they speak regularly.
Other demographic items were included to describe who visualizers are as people. We asked individuals to report their gender, including an option for “Prefer to self-describe.” Individuals were also provided the opportunity to identify as LGBTQ+ and if they consider themselves to be a member of a historically disadvantaged or underrepresented racial or ethnic group. Race or ethnicity were not asked in this survey as these characteristics vary widely across the globe and responses could present challenges with maintaining anonymity.
A small number of survey items were deemed sensitive and accessible only by the DVS board and survey organizing committee. These include fields with personally identifiable information (known as PII): locations more granular than country, language(s) spoken, job title, LGBTQ+ identification, and identification as a member of historically disadvantaged or underrepresented groups. Some of these items are shown in aggregate in this report; none were included in the full dataset released to the public for analysis and visualization.
What roles do we occupy, how much do we make, and how do we spend our time?
We answered the second question with responses about data visualizers’ occupations. Though many are not in roles specific to data visualization, many do work that incorporates some data visualization. To understand the spectrum of occupations, we asked for job titles; responses to this question were cleaned for summarization and are only shared with the public in a reduced dataset. Individuals who work in an organization or selected Freelance/Consultant/Independent contractor when asked to describe their role were also asked to identify the descriptor of the role closest to their job functions. These included analyst, scientist, engineer, developer, journalist, designer, cartographer, teacher, or leadership. Individuals within an organization were also asked to identify the industry (or industries) in which the organization operates.
We included a few questions about annual and hourly compensation, converted to United States Dollars (USD) for ease of comparison. Even though discussions of money can be considered impolite in some cultures, we felt it is important to share among peers or with those looking to make data visualization a more significant part of their livelihood to inform expectations and encourage everyone to appropriately value their own and others’ work.
The final part of this question is answered by individuals sharing how much time they devoted to data visualization activities in their most recent work week. The question was worded this way, rather than reporting about their average work week, as people provide more accurate estimates of recent events (e.g., Clarke, Fiebig, & Gerdtham, 2008). Even if some individuals had atypical work weeks, across respondents, work weeks are likely atypical in different directions (e.g., some people spent less time on their typical tasks, others spent more time on their typical tasks) and the aggregate responses are still informative.
How do we visualize and what challenges do we have?
A series of items in the survey focused on tools and technologies, popular visualization types, and challenges and frustrations experienced when they visualize data. Visualizers indicated what technologies they use often to visualize data. The response options ranged from platforms for mapping to data analysis to graphic design, from coding-heavy to point-and-click, and from dashboard to pen-and-paper drawing. Freelancers and those with a role in an organization also shared what problems they encounter when their tools are chosen by others. Response options were informed by previous surveys and also had space for visualizers to report new challenges.
As another way to understand how individuals are visualizing, they shared examples of how they communicate their data visualizations (e.g., dashboard, presentations, apps, physical media). Individuals also shared a broader perspective on their top three biggest frustrations when it comes to data visualization. Similar to the problems question described above, response options as well as a space to specify other frustrations were provided.
How has the COVID-19 pandemic impacted data visualizers?
Rounding out a second year of a global pandemic, we added two questions about how visualizers might have been impacted. First, individuals were asked if they changed employment or employment status in the last year and following that question were asked if the COVID pandemic and related economic crisis was a factor in their employment status. It was asked this way as people may have changed employment based on factors related to the pandemic or they may have retained their current status and decided not to make a change due to the pandemic context. Second, individuals shared if the pandemic has negatively affected their compensation or benefits.
What do the data tell us about the future of the industry?
Many questions can inform both the current and future state of the data visualization industry. We can show how many visualizers have changed employment or are looking for new opportunities. Distributions of years of data visualization experience across all five years of survey data provide an opportunity to look into the past to inform our future. Visualizers who have 5 years of experience or less were also asked what barriers they have faced in getting into data visualization professionally. Looking forward, visualizers of all experience levels shared their next priority for learning and preferred learning methods for new data visualization skills. Individuals also shared what they would be likely to use professional development funds on, providing information on what resources are highly valued.
What else should we know about the state of the data visualization industry?
Finally, to round out the findings, individuals identified their one to three most important issues they believe are facing data visualization. Response options were informed by previous survey items as well as popular topics among data visualization practitioners. Visualizers were also provided the opportunity to describe what others just don’t get about their visualization work. Themes emerged from the data (inductive) and were coded.
Survey Distribution and Sampling
The 2021 SOTI survey was disseminated online between 01 September and 01 October 2021. Distribution mechanisms included an email announcement and series of reminders from the official DVS email address, an announcement in Nightingale (the DVS publication), a post on the American Evaluation Association blog (AEA365), a series of posts on the DVS Slack, and numerous social media posts (e.g., Twitter, LinkedIn) from DVS and others (e.g., Meetup Groups).
Survey data were cleaned prior to analysis. Cleaning included standardizing open responses for monetary units, cities, and spoken languages, standardizing response options to align with surveys across years, and removing personally identifiable information (e.g., websites, unique roles) from open response entries. The design of the survey prioritized multiple choice (predefined) responses wherever possible, so fewer questions required cleaning than in past years. This tradeoff aimed for better per-question response rates (i.e., higher n) and less messy data, but at the expense of less granular data (e.g., no continuous numeric responses resulted in the inability to calculate means for many variables).
A total of 2,182 responses were submitted; 17 of these were removed from the final count because they only answered the first question and provided no other usable data. Survey responses from all five years were checked for accidental duplicates within the same year; however, because the data were anonymous, there was no way to check for double submissions from the same individuals if they selected different answers in the same year’s survey.
Survey results are primarily presented descriptively, focusing on group sizes (ns) and percentages. Because the survey was programmed with different branches and other mechanisms to reduce the number of items for each respondent, denominators for each question also vary. These are included in the narrative interpretation or in figures of relevant questions in the Findings section.
Basic qualitative analyses were conducted using both inductive and deductive approaches. Where qualitative responses were in an “Other, please specify” format, responses were first scanned to determine if they fit into any of the predefined response categories or anticipated response groupings (deductive). Responses that did not fit into these categories, like responses to purely free-input questions, were reviewed and categorized as themes emerged (inductive).
Suggestions for Further Exploration
Developing our understanding of data visualizers today should not stop here. This report offers only a start for learning from the 2021 Data Visualization Society’s SOTI survey results. From here, we call upon our fellow visualizers, both survey respondents and onlookers, to take this rich dataset and dissect, analyze, and showcase who we are and what we’re made of. We will leave you with a few parting ideas for further exploration of this and other data visualization data.
We found that most respondents want more time, but we don’t know how those who want more time might be different from those who did not indicate that time was a barrier or frustration. While many top barriers and frustrations are unsurprising, what is interesting is that there are substantial segments of respondents who did not find these top issues as problematic. How do they differ? Are they in certain sectors or do they come from a certain educational background? Do they make visualizations in certain ways or for certain people? The answers to these questions can give insight to the recommendations and resources we provide to those who are facing certain challenges.
We saw that roles and titles of respondents may be shifting to focus on visualization, but we don’t know how job titles intersect with specific roles, or if individuals with similar titles spend equivalent amounts of time on data visualization tasks. For example, what job titles do leaders have? How does the time spent on data visualization tasks differ between data visualization specialists, designers, consultants, engineers, and developers? Additional analyses and systematic parsing of the job titles field would help inform who we are and how we work in our respective domains, providing insight for individuals as they start on their data visualization journey or those seeking to understand data visualization work from positions of leadership or other visualization-adjacent roles.
We know that respondents have a range of data visualization and professional experience, but we don’t know how characteristics might differ between newer visualizers and more seasoned professionals. Do they use different tools or design different visualizations? Do they occupy similar roles, or do we see changes in role types as someone has spent longer in the industry? This type of evolution through experience and careers will be useful to track over time.
We explored many questions segmented by individuals who identified as doing freelance work, consulting, or work as an independent contractor as well as those who hold a position in an organization, but we don’t know a lot about hobbyists, students, academics, or those who earn a passive income from data visualization. Do they have specific barriers or challenges? Who do they make visualizations for? One report cannot do justice to all our unique roles within and contributions to data visualization, and it is important to learn what we can and share insights from these distinctive groups.
We showed how respondent countries, data visualization experience, and top tools changed over time, but we don’t know how other variables have evolved – or not. Does the popularity of chart types change over time? What about barriers and challenges? Do any of these changes differ by role or industry? Comparing this year’s SOTI survey data with prior years’ survey responses can provide a rich understanding of the patterns and ongoing or emerging needs of data visualizers.
At the beginning of the report we summarized other recent surveys of visualizers, but we don’t know how our respondents compare. The Observable and Viz for Social Good surveys had different purposes, but shared many similar demographic and descriptive questions with the DVS SOTI survey. Are each of these surveys telling the same story about who data visualizers are? If not, what are the differences and what other questions can we explore to understand these differences? Building synchrony across organizations to understand our overlapping interests and challenges could bring more resources to support data visualizers worldwide.
These data provide a mere snapshot of who we are as visualizers at this moment in history, though only time will tell how we evolve to merge our skills, our passion for learning, and our attention to the important issues facing data visualization today to shape and evolve the data visualization industry.
The DVS SOTI survey is created, managed, analyzed, and published with support from the DVS Survey Committee including Amy Cesal, Alexander Furnas, Josephine K. Dru, Lisa Valade-DeMelo, Sheila B. Robinson, and Jill A. Brown. Josephine Dru completed the initial data cleaning, Jill A. Brown was the primary author of this report, with significant input from Josephine K. Dru, Erica Gunn, and Lisa Valade-DeMelo.
Clarke, P. M., Fiebig, D. G., & Gerdtham, U. G. (2008). Optimal recall length in survey design. Journal of Health Economics, 27(5), 1275-1284. doi:10.1016/j.jhealeco.2008.05.012
Mauri, M., Elli, T., Caviglia, G., Uboldi, G., & Azzi, M. (2017). RAWGraphs: A Visualisation Platform to Create Open Outputs. In Proceedings of the 12th Biannual Conference on Italian SIGCHI Chapter (p. 28:1–28:5). New York, NY, USA: ACM. https://doi.org/10.1145/3125571.3125585
Wee People Font Copyright (c) 2018 ProPublica and Alberto Cairo.