Ethics in data visualisation
Data scientists are often the go-to people in an organisation to get a sense check on the organisation’s performance metrics. Therefore, they have a responsibility to portray the figures accurately. To achieve this, data scientists must adopt responsible visualisations and storytelling practices, void of bias. Results that are directly or indirectly manipulated due to personal bias, or to fit desired organisational outcomes, ideologies, or agendas, devalue the overall practice of data science and compromise its future.
Understanding your ethical position
A successful data scientist adopts responsible data-visualisation practices and maintains high ethical standards for themself and their work. You can learn to do this by understanding your ethical position in the context of your professional role.
When you’re faced with an ethical decision about data, consider the technical constraints, the legal constraints, and the organisational constraints.
Example
Your organisation has customer data that includes personal information from over a decade ago. As an ethical scientist, you know that this data was collected before the General Data Protection Regulations (GDPR) were enacted, and customers had no opportunity to decide about the use of their personal data. Now your manager is asking you to analyse this data and they want you to connect it to publicly available social media data.
Technically, this can be done – you know it can. The organisation is asking you to do it because it will serve the bottom line but, legally, the waters are a bit ethically murky. According to GDPR, an organisation needs to deal with historical data in one of three ways.
They could:
- delete the historical data
- devalue the historical data (make it ‘unlinkable’ or de-identified)
- transform the historical data to make it legal. [1]
If your organisation is holding this data in its original form without consent, they are acting illegally. This means that your ethical position will be the deciding factor for analysing the data.
A critical lens
Many data-related decisions are ambiguous. You’ll need to think critically and ask challenging questions in your organisation every step of the way.
Learning to read and interpret visualisations through a critical lens will improve how you create visualisations.
We’re immersed in a world of graphics and visualisations. Next time you encounter a data visualisation, take some time to look at it through a critical lens and ask yourself these three questions:
- What is the visualisation telling you?
- What is the motivation for the visualisation?
- What has been left out?
Answering these questions will help you to draw unbiased conclusions from data visualisations.
Example
Imagine you share the following visualisation with a colleague.
Your colleague says that your choice of colours suggests that Kenneth, Barney, and Aziz are somehow connected to the higher revenue. She also points out that your choice of colour suggests a binary gender bias. How would you respond to this feedback? What would you learn from this exchange?
Commit to your own ethical guidelines
An ethical data scientist is prepared to face the uncertainties and challenges of data by starting with their own set of ethical guidelines.
Here are some you may want to consider:
- I will respect the person that owns the data. I will treat their data as if it were my own. I will not misuse, mistreat, or exploit it.
- I will build trust with my audience by owning up to my mistakes and continually striving to improve.
- I will present data with honesty and integrity. The numbers don’t lie. I will check my work for accuracy but I will not hide the results.
- I will not invent fantasies about my data and I will accept that, sometimes, the data will be bland and predictable.
- I will know my data. I will understand it so that when I am communicating the findings – visually or otherwise – I will not mislead unintentionally.
- I will practise looking at data through a critical lens.
- I will share what I know about data with others when it could affect them.
Most importantly, don’t allow data to seduce you into acting unethically, and take your role as an ethical data scientist seriously.
References
- Lafever G. How to continue lawfully using historical data under the GDPR [Internet]. Anonos; October 2017. Available from: https://www.anonos.com/dont-lose-acccess-to-your-historical-data-under-the-gdpr
Introduction to Data Science for Business
Register to receive updates
-
Create an account to receive our newsletter, course recommendations and promotions.
Register for free









