The Session That Didn’t Want to Die #
In 2022 I wrote a blog post called “Retiring a Session” that talked about me retiring my most successful session ever. I had every intention of not running it again in the form it had, but it seems it is a session that just won’t die without a fight. I’ve since run it three times more, and it is on the schedule for the Power BI Gebruikersdagen in Utrecht, the Netherlands, in 2025.
But as I wrote in my previous blog post, the idea is not to kill it. Far from it. The world has changed, and while the message about data literacy is more important than ever, the way I go about discussing it and the examples I take need to be tweaked.
Why “Soft” Skills Need Constant Updates Too #
I find it fascinating that such a “soft” concept as data literacy would require constant updating just like a technical session would, but it goes to show that everything changes. Not only the technology.
The last few years have seen an erosion of trust in democratic institutions and the weaponization of data in political discourse. Likewise, I don’t think anyone could have predicted the huge impact of LLMs in everyday life and the disastrous effect on critical thinking and data/AI literacy this combination has had. When AI tools confidently present statistical hallucinations (and I hate the term “hallucinations” as it implies agency; more about that in a future blog post) as facts, and social media algorithms amplify misinformation faster than fact-checkers can respond, we face a crisis of data literacy unlike anything we’ve seen before.
It is more important than ever to combat this, and without technical professionals pointing out the dangers, we risk being drowned by the narrative that AI will solve our every need.
Introducing the Spiritual Successor: “Invisible Insights” #
I’ve been thinking about a spiritual successor to “The Untruthful Art” and I am very happy to say I have finally crafted the first draft. It will be called “Invisible Insights: What Your Data Can’t Tell You” and will focus on data analysis, explaining the need for domain knowledge, and why patterns in the data might not show what you think they do.
While “The Untruthful Art” focuses on how data is manipulated to deceive, “Invisible Insights” will examine the equally dangerous problem of what we fail to see: the biases we bring, the context we ignore, and the patterns we over-interpret.
I’m currently building it and I will be starting to submit it as soon as I have a solid draft. I have learned my lesson from writing abstracts and getting them accepted far before they were even conceptually done. Not doing that again!
Revamping “The Untruthful Art” #
At the same time, I’ve been giving “The Untruthful Art” a lot of thought as well. I’ll be tearing out some parts that don’t really align with the story, and I will move away from the somewhat disjointed five ways of misrepresenting data. They will still be there (they still need to be discussed and explained), but they will move around a bit and fit better in an overarching story.
I’m adding a section on AI-generated information and tying that into AI/data literacy, as well as giving all the examples a do-over.
It’s Time to Talk About COVID Data #
When I initially revamped this session during the pandemic, I deliberately avoided COVID data. Not only were (and still is) the topic very sensitive, but the data was at best incomplete and at worst completely wrong. Now, with more objective and qualitative data, it’s time to examine not just what the data told us, but how it was weaponized, misinterpreted, and used to further competing agendas. From manipulated axis scales on state dashboards to the deliberate confusion between daily cases and cumulative counts, the pandemic provided a masterclass in data deception that we cannot ignore.
New Topics and Modern Examples #
The updated session will include contemporary examples that resonate with today’s challenges:
- AI benchmark gaming: how companies optimize for metrics that don’t reflect real-world performance
- Social media misinformation dynamics: why 59% of links shared on social media are never clicked, and how this amplifies false narratives
- Election polling misrepresentation: recent examples of how the same data can tell completely different stories
- Climate data manipulation: examining both denial tactics and overcorrection in climate visualizations
Our Responsibility as Technical Professionals #
The more I work with literacy in general and AI/data literacy in particular, the more I feel that I need to do more in this space. As technical professionals we have a duty to try to stem the deluge of crap that floods the data space every day. We need more of us to stand up to misconceptions, misrepresentations and outright lies. Our livelihood, our value, our very relevance depends on it.
The words of Teller of Penn and Teller still ring clear in my mind:
“Nothing fools you better than the lie you tell yourself.”
We all see what we want to see. Let’s help our clients see what they need to see.
Join the Conversation #
Do you have a specific topic that you think should fit in the revamped session? I’d love to hear what examples of data deception or misrepresentation you’ve encountered in your work. Please reach out to me or comment on LinkedIn or BlueSky!
The revamped “Untruthful Art” will debut at Power BI Gebruikersdagen 2025 in Utrecht.