Why Most Data Visualizations Fail
And How to Fix Them (with Matt Harrison)
This week on the Data Neighbor Podcast, we’re diving deep into one of the most underrated superpowers in data science: communication through effective visualization. Our guest this week is Matt Harrison, bestselling author of Effective Pandas and Effective Visualization, Python educator, and a true master of practical data storytelling.
Matt shares why simplifying, not dazzling, is what actually influences stakeholders -and why tools like matplotlib are still king when it comes to storytelling with data. We talk about his framework, CLEAR, which lays out exactly how to elevate your visualizations from generic charts to persuasive, stakeholder-ready narratives. Whether you’re just starting out or you're a seasoned analyst, this episode offers insights you can immediately apply to your next project.
Why communication is the #1 skill for data professionals.
The 5 plot types most professionals use and why fancy plots usually backfire.
How to use color to direct attention and not distract.
Why so many visualization tools get it wrong and how to iterate like a pro.
The CLEAR method: a practical approach to crafting meaningful visuals.
Key Takeaways
Focus on the Story, Not the Tool: Tools like Tableau and Plotly are great, but they often restrict your ability to refine and customize for clarity. Python with Matplotlib gives you full control to tell the story you need to.
Use CLEAR to Guide Your Visuals: Matt’s framework - Color, Limited plots, Explain with titles/subtitles, Audience, References - acts as a checklist to ensure your visualization is actually effective.
Practice Like a Designer: The best way to improve is by recreating great visualizations you admire, paying attention to color, layout, and storytelling.
Know Your Audience: Not every chart needs to impress a data scientist. Tailor your choices to what the decision-maker needs to understand - quickly and clearly.
Beware of Defaults: Defaults (like 10-bin histograms or spaghetti line plots) often mislead. Learn to intentionally adjust elements to suit your narrative.
Links Mentioned
Matt's site: https://www.metasnake.com
Chapters
0:00 Why Visualization Matters
1:43 Meet Matt Harrison
3:12 Why He Wrote Effective Visualization
6:30 Why Matplotlib Still Wins
9:28 Why AI Isn’t There Yet
15:33 Introducing the CLEAR Framework
16:26 C - Color: Be Intentional
17:54 L - Limited Plots: Use the Basics
21:15 E - Explain: Title, Subtitle, Annotations
23:08 A - Audience: Who Are You Speaking To?
24:52 R - References: Establish Authority
31:04 Iterative Process in Practice
34:22 How to Know If a Plot Is "Good"
42:20 Common Mistakes in Visualization
48:06 Can LLMs Improve Charting?
56:14 Most Underrated Skill: Communication
58:46 Where to Find Matt Online
Connect with Hai, Sravya, and Shane:
#datascience #datavisualization #python #matplotlib #effectivecommunication #analytics #storytellingwithdata #matthewharrison #dataanalysis #dataeducation #careergrowth #dataneighborpodcast #clearmethod #machinelearning #ai
What a great, insightful talk! I may not agree with everything, but it’s always enjoyable to hear great data scientists sharing their perspectives on data visualisation, etc.! Love it.