Equitable methods for data collection and information visualization.
By Christina Singer
Designing equitable information and visualizations involves self-examination, attention to others, and intentionality throughout the design process. You are possibly a user of your product, but you are certainly not every user. No amount of empathizing with others will change anyone’s circumstance—learning from and designing with others can, though. Humans in decision-making positions must be reflexive and expand data collection and visualization mechanisms to improve the quality of datasets and achieve better experiences. This article shares three actionable strategies that can be used to activate qualitative data and build equitable visualizations and solutions [see Figure 1].
Extract Yourself
Positionality
Positionality is a combination of factors in your life that shape your way of seeing the world and inform your unique position. Your positionality affects how you communicate, make decisions, and navigate life. To start, I will visually communicate my positionality [see Figure 2].
By mapping and reflecting on your positionality, you are able to better identify and articulate your own biases, privileges, and potential blind spots when trying to empathize with others. You have experienced life one way, which has given you your own way of seeing the world and approaching circumstances. Therefore, you can’t truly relate or empathize with the exact experience someone else has had.
Mapping or clustering your stakeholders’ positionalities can help you better understand their circumstances and where they have come from in life when they approach your product or service. It is important for teams to understand who the users are and what they need or know by gathering qualitative data before or with quantitative data, and then reflecting on the results of both data sets. The list below [see Figure 3] is composed of various factors to consider when forming user interviews or surveys to determine the intricate positionalities of your users, and in turn inform your visual and written communications to them.
The list below is a starting point, as individuals will likely come up with other factors to explain their unique positionalities related to your product or project. Note that many of the factors—which impact our lives from childhood and inform how we may inherently see and interact with the world—are out of our control. These factors inform how we make decisions, what knowledge we have access to, and how we respond to language and symbols used in written and visual communications.
Lesley-Ann Noel and Marcelo Paiva (2020, February) published a publicly-available online article “Learning to Recognize Exclusion,” from the Journal of Usability Studies, which includes a unique way to visualize and cluster positionality maps from a group of people into a “positionality radar chart” using a worksheet they designed to help teams determine equity gaps. The worksheet focuses on eleven categories, and is available to download from the references section of their paper, or by searching “positionality radar” in the Figma Community. This resource is a useful tool that can be used to visualize and identify equity gaps in a group of people, such as your team or your users. When you have gathered and visualized the data from your users or team, that is just the starting point for reflection and action.
When learning about others’ positionality, it is important to avoid a deficit mindset. In the book From Equity Talk to Equity Walk, Tia Brown McNair, et al. (2019) explore how deficit-mindedness can lead to cold, demeaning language being used to communicate in documents. Look at your positionality map and look at someone else’s. What are you assuming about them? If you naturally view a difference in a lived experience between you and them as negative, then you are assuming a deficit mindset. Learning more about your positionality and consciously unlearning that differences with others are negative are two good steps towards creating more equitable data collection and human-centered solutions.
Critique Your Process
What methods does your team use throughout your creative and research process, and how can those methods be biased or become more equitable? For example, personas are a common method used in design processes, and are typically informed by primary and secondary research. Your team may synthesize and simplify information about a group of users, and use that data to generate a persona that represents the goals, values, and frustrations of a larger group of users throughout the design process.
With the knowledge that everyone’s positionality is so dramatically different, how can you be conscious of this when creating and using personas? Another great equity resource from Noel (2020, October) is “The Designer’s Critical Alphabet.” The website, criticalalphabet.com, poses terms such as Westernization and xenophobia, and the user is encouraged to question if their design has considered these critical terms and reveal any blind spots or biases. I use this list of terms with my students in a senior-level design research course to critique methods used by IDEO, IBM, and other companies. We discuss how we could use or adapt certain methods to be more mindful and equitable. How might your team apply these terms to critique your process?
Embrace Others
Co-design
A particular usability testing method from IDEO that stands out as potentially problematic is the “empathy tools” method (William Stout Architectural Books, 2003). This method attempts to simulate an impairment of the user when testing a product to determine the usability of the product and the user’s needs. The designers would simulate a partially blind user with foggy goggles or a user with limited mobility in their hands with clunky gloves. This simulation is problematic for a number of reasons, but the primary reason is that no simulation is ever comparable to the real experience. In designing a product for a group of users, the design process should involve the users as co-designers without making assumptions or trying to simulate the user’s experience. This is called co-designing. Many methods from IDEO and other organizations with similar design thinking frameworks do embrace co-design in a range of methods. It is important for teams to address bias throughout their process in order to reveal their limitations and disclose their inability to truly empathize with certain users.
Per Mollerup (2015) summarizes Steven Kosslyn’s eight psychological principles for creating successful data visualizations, and the “principle of appropriate knowledge” stands out when co-designing, such that the creator of information or visual designs should use appropriate language for users and not patronize them or use unrelatable industry jargon. Kaleena Sales (2022) notes in her essay in the book The Black Experience in Design that when we prioritize communicating to the audience, we should consider the cultural backgrounds of the primary users and include pertinent visual language while steering away from the International Typographic Style as a default. What would it look like to make the visual language of your product more culturally aware and accessible to the primary users of your product?
Accessibility
Accessibility is not only done in good faith to create designs that work better for everyone, but it is also a legal obligation for companies to provide equitable digital and physical experiences. The government has set standards in terms of what qualifies as accessible for the web, but those standards don’t necessarily come with directions for how to achieve equitable products. AccessiBe and other online services that work to make your products more accessible can be great, but pricey. IBM Accessibility is a free, equal access toolkit that provides directions for creating accessible user experiences.
While some simulations of experiencing impairments fall short, such as empathy tools, others can be more accurate for creating products to then test with users experiencing certain impairments. One particularly successful free accessibility checker, Color Oracle, simulates colorblindness by changing how you see color on your computer screen. It is compatible with Windows, Mac, and Linux. Another tool that checks for legally-necessary color contrast in interfaces, typographic hierarchy, and allows you to create flows that direct e-readers throughout the page is Stark, a software plugin for Figma, Sketch, Adobe XD, and other programs. Stark has a limited free base plan, but to access more features you will need to pay.
Semiotics
It is critical for the designer and developer team to avoid stereotyping any culture or subculture with color palettes, iconography, terms, and other imagery and graphic elements when storytelling with data or information. One of the best ways to know if elements of your design would be offensive to members from a diverse audience is to have designers from a range of subcultural backgrounds represented on the team. If that is not currently the case for your team, and a large portion of your users or viewers are unlike yourself, then it is your duty to thoroughly research the meaning of imagery you are using and test the product with a diverse group of users. Your error could be as seemingly simple as using a culturally insensitive typeface. Ellen Lupton (2021) writes about how people often stereotype cultures as “tribal” when they misappropriate typefaces to represent a culture other than that which the typeface was originally designed to symbolize.
As Ruben Pater (2018) writes about cultural differences in The Politics of Design, visual literacy has to be learned, and it develops in different ways across cultures. Transcreation is the process of translating stories and visuals from one language and culture to another while maintaining the integrity of the original concept. Illustrators, copywriters, artists, and designers who are launching products across cultures should practice transcreation by commissioning illustrators and translators from diverse cultures to adapt the visuals and language for their cultures. For instance, when the Powerpuff Girls entered Japanese markets, the anime underwent visual and linguistic transcreation through a collaborative process between Cartoon Network and Aniplex to better suit the Japanese audience. Similarly, sometimes culturally-aware subtle shifts occur in films, such as Inside Out and Big Hero 6. In Inside Out, the vegetables on the child’s plate change from broccoli to peppers based on the viewer’s location. As for Big Hero 6, the Japanese names of the main characters change for Korean audiences due to deeply-rooted historical tensions between the countries.
The use of terms and imagery that seem inoffensive to you might trigger generational trauma for someone else. Adobe InDesign recently renamed master pages to be called parent pages—a direct response to the connotation that the term “master” has in the historical context of slavery. Many industries are rethinking terms and even brands by revisiting their use of offensive language and visual imagery and making changes to terms and imagery in response to negative associations with those words or symbols. How might your team analyze your work and consider being more mindful of how various users could interpret your language or visuals based on historical connotations?
Expand Data Collection and Visualization Mechanisms
Collection
While data collection can never be purely objective, it can do better. User testing groups for products with broad audiences should be intentionally diverse, minding qualitative data derived from the unique positionalities of individuals [see Figure 3]. Through the Algorithmic Justice League (AJL.org), computer scientist Joy Buolamwini—aka Dr. Justice—talks about poorly engineered products that lead to discriminatory user experiences due to inequitable data collection processes. Machines and artificial intelligence that detect voices and faces, as Buolamwini notes, are created by humans, and datasets that inform them inherit the biases of their makers. Inadequate diversity in user testing groups is part of the problem.
As Catherine D’Ignazio and Lauren F. Klein (2020) point out in their book Data Feminism, it matters what we count, such that even quantitative datasets are often incomplete or misinformed due to the exclusion of nonbinary data collection. Most large institutions, such as universities, are still only collecting data on users based on gender binaries, which leads to inaccurate representations of human experiences and solutions that don’t account for those excluded from the system that was designed based on inequitable data collection.
Visualization
Giorgia Lupi (2023, February) coined the term data humanism to describe a more personalized and human data collection and visualization process that embraces visualizing qualitative data such as human stories. Data represents humans, but often lacks any human quality beyond numerical interpretations of reality. Qualitative data can make quantitative information more usable and informed, leading to more human-centered solutions. Collecting and visualizing both qualitative and quantitative data is often necessary in order to most accurately understand the nuanced details of the humans and systems being evaluated. What if your exercise app asked and collected why you exercised more or less that day, and offered you health tips to address your reality rather than simply quantifying and visualizing your highs and lows? Lupi asks us to consider the following question when reflecting on the data that we have collected: “What can we learn from this on a more human level—and what does it mean to the world?”
Reflection
Data that is collected should be accurately represented in a visualization, but our responses and problems we identify when synthesizing the results to propose solutions or changes must take into account the human narrative underneath the numbers. Mindful reflection on the data collected is required before it is visualized, and then again after it is visualized (McNair et al., 2019). Where do you observe differences in the data for various users, how might those differences be caused by inequities, and why? What additional qualitative and quantitative data might you need to collect in response to your new questions from reflecting on your visualization? Equitable solutions aren’t natural in a world that has been designed to benefit some more than others, so as an actor in this system, how can you do better?
References
D’Ignazio, Catherine, & Klein, Lauren F. Data Feminism. The MIT Press, 2020.
Lupi, Giorgia. “Data Humanism: The Revolutionary Future of Data Visualization.” Print magazine. Last modified February 25, 2023, https://www.printmag.com/article/data-humanism-future-of-data-visualization/
Lupton, Ellen. “Giving and Taking Credit.” Extra Bold: A Feminist, Inclusive, Anti-racist, Nonbinary Field Guide for Graphic Designers. Princeton Architectural Press, 2021, 196-197.
McNair, Tia Brown, Bensimon, Estela Mara, & Malcom-Piqueux, Lindsey. From Equity Talk to Equity Walk. 1st ed. Wiley, 2019.
Mollerup, Per. Data Design. Bloomsbury, 2015.
Noel, Lesley-Ann. Critical Alphabet. Last modified October 4, 2020, http://criticalalphabet.com.
Noel, Lesley-Ann, & Paiva, Marcelo. “Learning to Recognize Exclusion.” Journal of Usability Studies 16, no. 2. (February 2020): 63–72. https://www.researchgate.net/publication/349642679_Learning_to_Recognize_Exclusion.
Pater, Ruben. The Politics of Design: A (Not So) Global Manual for Visual Communication. BIS Publishers, 2018.
Sales, Kaleena. “Beyond the Universal: Positionality & Promise in an HBCU Classroom.” The Black Experience in Design: Identity, Expression, & Reflection. Allworth Press, 2022, 170-175.
William Stout Architectural Books. IDEO Method Cards: 51 Ways to Inspire Design, 2003.
Christina Singer (csinger3@uncc.edu) is an Assistant Professor of Graphic Design at UNC Charlotte, where she teaches design research and UX/UI design strategies.
