Data Storytelling, AI
Redefining Data Storytelling with AI
A summary of the webinar I presented at the United Nations
Just analyzing and presenting data isn’t enough. To make an impact, you need to communicate your insights through storytelling. In this blog, I’ll share key takeaways from my webinar, Redefining Data Storytelling with AI, presented at the Global Network of the United Nations and attended by 243 participants worldwide. Let’s explore how AI is changing the way we tell data-driven stories.
If you prefer, you can watch the video recording directly.
What is Data Storytelling?
At its core, there are three types of data communication:
- Data Reporting — Describing the data.
- Data Presentation — Organizing it effectively.
- Data Storytelling — Crafting a compelling story that resonates.
Data storytelling is all about making your data meaningful. It’s not just numbers and charts — it’s about engaging your audience, informing them, and inspiring action.
Understanding Your Audience
Every great story is tailored to its audience. Who are you speaking to?
- General Public — They want to be informed or entertained.
- Professionals — They need details and insights.
- Executives — They want clear takeaways for decision-making.
To connect with your audience, you need to adjust your language, tone, and level of detail. Context matters!
Language refers to the choice of words and the complexity of your phrasing. Use simple, everyday words and avoid technical jargon when speaking to a general audience. However, for professionals or executives, you may incorporate more specialized terminology and concise, data-driven statements.
Tone conveys the emotional and rhetorical style of the message. For instance, an educational tone might be used for the general public, an authoritative and confident tone for executives making decisions, and an analytical tone for professionals.
The right balance of language and tone ensures that your audience can connect with the story and take action based on the insights presented.
The Three-Act Structure of Data Storytelling
A strong data story follows a classic three-act structure:
- Context — What’s the background and why does it matter?
- Main Point — What’s the key insight from the data?
- Next Steps — What actions should be taken based on this insight?
While the main point stays the same, you should tweak the context and next steps based on your audience.
How AI Enhances Data Storytelling
AI makes data storytelling more engaging and audience-specific. Here’s how:
1. Tailoring the Story to the Audience
AI helps adjust language, tone, and content for different readers. To tailor a story for a specific audience, you need to structure your prompt effectively. A well-structured prompt should include:
- Audience Definition: Specify who the story is for (e.g., general public, professionals, executives).
- Purpose: Define the goal of the story — whether it’s to inform, persuade, or entertain.
- Tone & Language: Indicate whether the story should be formal, conversational, technical, or simplified.
- Context & Example: Provide a specific scenario or analogy that resonates with the audience.
Incorporating these elements makes AI-generated content more engaging and appropriate for the target audience.
For example, consider the following simple visual data story.
You can adapt the title and the commentary for different audiences:
- For the General Public:
- For an Italian Audience (using a food analogy):
- For Professionals:
2. Evaluating Story Quality
AI can assess whether a story is clear, engaging, and relevant. But I’m still digging deeper into how best to measure AI-generated storytelling. How do we ensure AI-crafted stories are meaningful and unbiased? This is an exciting challenge I’m currently exploring.
3. Generating Next Steps
AI can also help suggest actionable insights. Take the issue of rising global temperatures — how should different audiences respond?
- For Decision-Makers:
- For the General Public:
- For Climate Skeptics:
The Limitations of AI in Data Storytelling
AI is powerful, but it’s not perfect. Here’s what you need to watch out for:
- AI Hallucination — Sometimes, AI generates misleading or incorrect information.
- Bias in AI — AI can reinforce biases in data if not carefully managed.
- The Need for Human Oversight — AI-generated content should always be reviewed and refined by humans.
To address these challenges, I recommend fine-tuning AI models, using Retrieval Augmented Generation (RAG), and keeping humans in control of the storytelling process. Fine-tuning AI helps improve accuracy by training models on specific datasets to align with desired outputs. RAG enhances responses by integrating real-time data retrieval, making storytelling more precise and up-to-date. Additionally, human oversight ensures that the generated content remains contextually relevant, ethical, and free from biases. By combining these strategies, we can harness AI’s potential while mitigating its limitations, creating powerful and responsible data narratives.
Conclusion
AI is transforming data storytelling, making it more personalized and impactful. But it’s not a replacement for human intelligence — it’s a tool that amplifies our ability to communicate insights effectively. If we use AI wisely, we can create stories that inform, engage, and drive action.
I’m excited to keep exploring how AI can enhance storytelling. Let’s push the boundaries of how we share data-driven insights!
