Staying ethical in data visualisation: Why, What and How?
Numbers don't lie, but the way we present them can. As our reliance on data-driven decision making grows, so does our responsibility to ensure these visualisations tell the whole truth. For data visualisation designers and developers, the primary focus is to create visuals that are eye-catching, simple, and accurately answer business questions by following various well-documented standard data visualisation practices.
However, ethical data visualisation goes beyond merely following standard practices. It emphasises accuracy, honesty, and clarity, ensuring that visual representations of data do not mislead or misinform. When combined with conventional data visualisation standards, ethical approaches produce more reliable and trustworthy data storytelling. This keeps audiences well-informed of the complete picture—whether good, bad, or ugly—ultimately supporting a more thorough decision-making process.
In this blog, we will delve into the full spectrum of ethical data visualisation, examining its definition, benefits, and guiding principles. We'll discuss how these principles apply to various stages of the visualisation process and look at examples of misleading and unethical data representations.
Do you want to take a step back and learn more about why data visualisation deserves the attention it gets? Check out our previous blog ‘3 key elements to get Data Storytelling right’.
Benefits of ethical data visualisation:
1. Enhances data accuracy & transparency:
By adhering to ethical standards, visualisers create clear, unbiased representations that truly reflect the underlying data. This approach prevents misinterpretation and allows audiences to understand the information's real meaning and context, leading to more informed discussions and decisions.
2. Boosts credibility and stakeholder trust:
When data visualisers prioritise ethics, they build trust with their audience. Ethical visualisations demonstrate integrity in data handling, respect for privacy, and commitment to fair representation. This builds credibility not just for the individual visualisations, but for the entire organisation or project. Stakeholders are more likely to rely on and value insights from sources they trust, enhancing the impact of the presented information.
3. Maximises impact of data & decisions:
Practising ethical considerations at each stage of the visualisation process (from data collection to final presentation) minimises biases and errors at each stage and ensures the delivery of reliable insights. This reliability is crucial for decision-makers who rely on these visuals to guide critical business decisions, ultimately leading to better-informed strategies and outcomes.
4. Fosters critical thinking & data literacy:
Ethical data visualisation goes beyond creating honest visuals; it also encourages critical thinking among viewers. By presenting data ethically, visualisers empower their audience to question, interpret, and understand the information presented. This also fosters improved data literacy, enabling viewers to better detect potential manipulation and make more informed judgments about the data they encounter in various contexts.
Now that we've explored the compelling benefits of ethical data visualisation, it's crucial to understand how to put these ideas into practice. Let's delve into the key principles that form the foundation of ethical data visualisation.
These principles provide a roadmap for creating impactful and trustworthy visual representations of data.
Principles of ethical data visualisation:
The 7 guiding principles of ethical data visualisation are 1. Accuracy, 2. Clarity, 3. Simplicity, 4. Objectivity, 5. Privacy, 6. Inclusivity, and 7. Accessibility. Here how to adhere to them:
Accuracy and honesty:
Being honest and accurate in data visualisation means presenting data truthfully, without distortion or manipulation.
This involves using the correct data, avoiding cherry-picking or misrepresenting information to support a specific viewpoint. To maintain accuracy, designers should verify data sources, cross-check information, and be transparent about data limitations or uncertainties.
Clarity and simplicity:
Clear and simple visualisations make complex data digestible and accessible. This can be achieved by choosing appropriate chart types, using consistent and readable fonts, and organising the layout for easy navigation.
Simplifying visuals helps the audience to quickly grasp key insights, patterns, or trends without confusion. To achieve clarity and simplicity, designers should prioritise essential information, minimise visual clutter, and use colours, labels, and legends effectively.
Fairness and objectivity:
Fair and objective data visualisations avoid biases and present data impartially. This principle ensures that visualisations don’t favour a particular viewpoint or mislead the audience with biased interpretations.
To stay objective and retain the fairness of data, designers should select unbiased data sources, acknowledge data limitations, and present alternative perspectives when appropriate.
Privacy and trust:
Respecting privacy and confidentiality in data visualisation means protecting sensitive or personal information.
This principle ensures that visualisations don’t violate privacy rights or expose confidential data. Respecting privacy and confidentiality involves anonymising data, aggregating information to a safe level, and obtaining consent when necessary.
To maintain privacy and confidentiality, designers should follow relevant data protection regulations, guidelines, and best practices throughout the visualisation process.
Inclusiveness and accessibility:
Cultural sensitivity and inclusivity in data visualisation require considering the diverse audience's needs, preferences, and backgrounds.
This principle ensures visualisations are accessible and respectful to people from various cultural and linguistic backgrounds, as well as those with different abilities. Implementing this may involve using appropriate colours, symbols, and language to avoid offending or excluding viewers.
Is your organisation currently facing challenges in implementing these ethical principles, need guidance on a specific project, or simply want to discuss best practices, our team of experts is just a message away. Get in touch today for tailored support.
Embedding ethics in every data visualisation stage
Ethical data collection
Data collection is the first stage in the data visualisation process, where relevant and accurate data is gathered from various sources. This essential step forms the foundation for subsequent analysis and visualisation.
Data can be collected from primary sources, such as surveys, interviews, or experiments. Also, it can be gathered from secondary sources, like databases, published research, or government reports. In some cases, data may also be obtained through real-time sources, such as sensors or social media platforms.
The two main ethical considerations in this phase are:
Ensuring data accuracy and completeness – Data collectors must strive to gather high-quality, reliable data to avoid misrepresentations or misleading conclusions. This may involve cross-checking sources, verifying data authenticity, or addressing potential gaps in the dataset.
Respecting data privacy and consent – When collecting personal or sensitive information, you must adhere to privacy regulations and get informed consent from data subjects. This involves being transparent about intent of data usage and safeguarding collected data to prevent unauthorised access or misuse.
Ethical data analysis
Data analysis involves examining and processing collected data to uncover trends, patterns, or insights that will inform the visualisation. This step is essential to transforming raw data into meaningful information.
For the analysis data visualisers use various techniques like descriptive statistics, data cleaning, data aggregation, or even more advanced analytics methods like machine learning. These techniques help reveal meaningful patterns or relationships within the data.
Ethical considerations during data analysis include:
Minimising biases and errors – Designers should use appropriate methods, tools, and techniques to reduce biases and errors during data processing. This involves critically evaluating data quality, being aware of potential pitfalls, and validating analytical results.
Transparency in data processing – Analysts need to be transparent about the steps, assumptions, and methodologies used during data analysis. This enables others to verify, replicate, or challenge the results, promoting accountability and trust in the findings.
Ethical Data visualisation (design)
Data visualisation is the final step in this journey. It involves creating visual representations of data that effectively communicates insights, patterns, or trends to an audience. Designers utilise various forms, such as charts, graphs, maps, or interactive elements, to make complex data more accessible and understandable.
Good design choices, including colours, scales, and layout, play a significant role in conveying the intended message and influencing viewer interpretation.
The main ethical considerations during data visualisation design include:
Honest data presentation – Designers must avoid manipulating or distorting data to mislead or misinform viewers. This involves choosing appropriate chart types, scales, and data transformations that accurately show the underlying data.
Accessible and inclusive design – Designers must prioritise creating visuals that cater to diverse audience needs and preferences, ensuring maximum accessibility of information. This inclusive approach involves considering various factors such as colour blindness, screen reader compatibility, and alternative text descriptions for visual elements.
To ensure cultural sensitivity and inclusivity in data visualisation, it's crucial to conduct thorough audience research and actively seek diverse feedback. However, this is just the beginning. Designers must also cultivate a mindset of openness and flexibility, being willing to adapt their work in response to varied perspectives and cultural nuances. This iterative process of research, feedback, and adjustment helps create visualisations that resonate with and respect a broad spectrum of cultural backgrounds and experiences.
We've journeyed through the ethical considerations at each stage of the data visualisation process, from data collection to final presentation. Now, it's time to apply these insights to your own work.
Are you facing a particular ethical challenge in your current visualisation project? Perhaps you're struggling with:
Ensuring data integrity during the collection phase
Choosing an appropriate visualization type that doesn't mislead
Designing for accessibility without compromising your message
Providing context that enhances understanding without bias
We're here to help you navigate these complexities. Send us a message at contact@eisquare.co.uk and outline your specific challenge. Our team will be in touch to guide you towards an ethical solution
Examples of misleading visuals due to unethical data visualisations
There are 6 common types of misleading visualisations: 1. Scale distortion, 2. Data not adding up, 3. Arbitrary dual Y-axes, 4. Manipulated Y axis, 5. 3D Distortion and 6. Omitted or concealed data.
Scale Distortion
Vertical axis not starting at zero or skipping numbers – truncated y-axis leads to distorted patterns and over-exaggerated differences, suggesting differences that do not exist. In most cases, the baseline should start at zero.
Data Not Adding Up:
When Graphs that don’t add up! Pie charts are used to show the parts of a whole, not the difference between groups. Pie charts should always add up to 100%.
Arbitrary Dual Y-Axes
The scales for these axes can be set arbitrarily, which may inadvertently or deliberately mislead readers about the relationship between two data series. For instance, in a graph comparing GPA (blue field) and units passed (red line), the scale manipulation can create visual distortions. Small differences in GPA might appear dramatically large, while significant variations in units passed could seem deceptively minor.
Manipulated Y-Axis
When the scale is disproportionate to the data, it can artificially minimise or exaggerate changes over time, leading to inaccurate interpretations. For instance, compressing the y-axis might make substantial fluctuations appear insignificant, while expanding it could make minor variations seem more dramatic than they are. To enhance clarity and promote accurate understanding, it's crucial to use appropriate scales and include clear labels showing actual values for each data point or time period.
3D Distortion
3D Graphs look cool and add dimension to an otherwise boring bar chart; however, they can distort the bars making differences appear larger compared to a 2D graphic. The depth perspective in 3D graphs can obscure true values, especially for bars in the background, potentially misleading viewers.
Omitted or Concealed Data
Omitted or concealed data can be either intentional (cherry picking) or unintentional. Researchers may only include data which reinforces their narrative. This can provide a false representation of the data
Data has the power to change the world – but only when it's wielded responsibly. As you leave this page, consider the ripple effect of your next visualisation. Will it clarify or confuse? Enlighten or mislead? . If you're feeling the weight of this responsibility and want to ensure your visualisations make a positive impact, we're here to help. Our team of visualisation experts is ready to guide you through your next project, big or small. Book a free discovery call with them using the link below.
Bottom line…
Ethical data visualisation stands as a cornerstone in conveying information honestly and responsibly, shaping perceptions, and influencing decision-making. It extends far beyond aesthetics, serving as a crucial tool for accurately conveying complex information and supporting informed decision-making.
Data visualisers and consumers can foster trust, promote responsible practices, and contribute to a more informed and fair society. We can do that by following good ethics in data visualisation.
By adhering to ethical principles in data visualisation, both creators and consumers contribute to a more transparent and trustworthy information ecosystem.