🎥 Watch the Full Workshop Series
📌 Part 1: https://youtu.be/ozOfDbbs69o
📌 Part 2: https://youtu.be/MltYW2QKJOY
Introduction
In today's data-driven world, effectively visualizing data is a critical skill for anyone presenting research, business reports, or scientific findings. However, poorly designed charts can mislead your audience, hide key insights, or create confusion. This guide, based on my two-part Better Data Visualization workshop, walks you through the best practices, common pitfalls to avoid, and effective techniques to create impactful data visualizations.
By the end of this article, you’ll know which chart types to use, when to use them, and how to optimize your visuals for clarity and accuracy.
🔹 Part 1: The Fundamentals of Data Visualization
1️⃣ Why Data Visualization Matters
Raw numbers in a table are difficult to process visually. Our brains process images faster than text, so an effective visualization allows your audience to:
✅ Identify patterns and trends 🔍
✅ Compare different data points easily 📊
✅ Avoid misinterpretations 🚫
✅ Focus on key insights quickly 🎯
For example, consider Anscombe’s Quartet—a set of four datasets with identical statistical properties (mean, variance, correlation), but when plotted, they reveal very different distributions. This highlights why visualization is essential for accurate data interpretation.
2️⃣ The Core Principles of Visual Perception
Our brains interpret visual information based on gestalt principles, which include:
✔ Proximity – Elements close together are perceived as related
✔ Similarity – Elements with similar color, shape, or size form a group
✔ Enclosure – Boundaries create distinct visual groups
✔ Closure – Our brains fill in missing information in a pattern
✔ Continuity – The eye follows smooth paths
✔ Connection – Lines connecting elements imply relationships
Using these principles intentionally in your charts can enhance readability and insight clarity.
3️⃣ How to Direct Audience Attention in Your Charts
✔ Use color strategically – Highlight key data points with color contrast, but don’t overdo it.
✔ Remove distractions – Unnecessary grid lines, legends, and excessive data points add clutter.
✔ Use labels wisely – Directly label important values instead of using a separate legend.
✔ Limit the number of colors – Too many colors create confusion instead of clarity.
📌 Example: Instead of an overwhelming color-filled bar chart, gray out less important data and only highlight the key trend.
4️⃣ Avoid These Common Mistakes in Data Visualization
❌ Never use 3D charts – They distort perception and make accurate comparisons difficult.
❌ Avoid spaghetti charts – Too many overlapping lines create clutter. Use small multiples instead.
❌ Reduce legend dependency – Directly label lines and bars where possible.
❌ Use effective titles – A title should summarize the key insight of the chart, not just describe the data.
❌ Don’t use dual y-axes – It can mislead interpretation by making unrelated data appear correlated.
5️⃣ Pie Charts: Should You Ever Use Them?
📌 Pie charts are problematic because humans struggle to compare angles and areas accurately. A bar chart is almost always a better alternative.
✅ When to use a pie chart: Only when comparing two or three categories, with a clear dominant value (e.g., “90% of cases are caused by E. coli”).
🚀 Better Alternatives:
✔ Bar Charts – Easier to compare values directly
✔ Tree Maps – Useful for hierarchical proportions
✔ Donut Charts – Only when emphasizing a single category
🔹 Part 2: Advanced Data Visualization Techniques
6️⃣ Visualizing Data Over Time
Best charts for time-series data:
✔ Line charts – The best for showing trends over time 📈
✔ Slope graphs – Great for comparing two time points 📊
✔ Sparklines – Small, inline charts for quick trend analysis 🔍
🚨 Common Pitfalls:
❌ Avoid using bar charts for time-series data – Line charts work better for trends.
❌ Avoid dual-axis line charts – They often mislead by forcing unrelated scales into comparison.
7️⃣ Showing Data Distribution
To analyze variability in data, use:
✔ Histograms – Show frequency distributions 📊
✔ Box plots (box-and-whisker plots) – Display median, quartiles, and outliers 📦
✔ Violin plots – Combine box plots with kernel density estimates 🎻
✔ Bee swarm plots – Show the exact distribution of individual data points 🐝
📌 Box plots are underutilized in research but extremely powerful for comparing multiple groups.
8️⃣ Visualizing Relationships Between Variables
For exploring correlations between two or more variables, use:
✔ Scatter plots – The go-to chart for relationship analysis 🔵
✔ Bubble charts – Add a third variable with bubble size 🔴
✔ Chord diagrams & Sankey diagrams – Show relationships between categories 🌐
🚨 Key Tip: Use color and size effectively in scatter plots to encode extra variables without overwhelming the audience.
9️⃣ Optimizing Data Tables for Readability
Not all data needs to be visualized—sometimes a well-designed table is the best option.
📌 Tips for effective tables:
✅ Bold headers and left-align text for readability
✅ Right-align numbers for easy comparison
✅ Reduce grid lines – Use subtle dividers instead
✅ Use color shading to guide the audience’s focus
✅ Add heatmaps or sparklines to highlight patterns
📌 Pro Tip: People naturally scan down a table—adjust spacing to guide them row by row, not column by column.
🔹 Key Takeaways & Best Practices
✔ Prioritize clarity – Choose the simplest chart that conveys your message
✔ Highlight key insights – Use color, labels, and layout intentionally
✔ Avoid 3D effects – They distort perception and reduce accuracy
✔ Minimize clutter – Every element should serve a purpose
✔ Choose the right chart type – Match the visualization to the data type
✔ Be intentional with tables – Guide the reader’s focus with structure
🚀 Final Rule: Make your data easy to understand, not just easy to create.
📌 Watch the Full Workshop Series
📽 Part 1: Fundamentals of Data Visualization → https://youtu.be/ozOfDbbs69o
📽 Part 2: Advanced Techniques & Best Practices → https://youtu.be/MltYW2QKJOY
💬 What’s your favorite chart type? Have you spotted bad data visualization in research or media? Drop a comment below! 🚀
🔔 Subscribe for more insights on data visualization, research communication, and effective presentations!
Alireza FakhriRavari, PharmD, BCPS, BCIDP, AAHIVP is Department Chair and Associate Professor of Infectious Diseases at Loma Linda University School of Pharmacy.
Comments