Simplifying Data with Visual Representations in Science

𝗙𝗼𝘂𝗿 𝗱𝗮𝘁𝗮𝘀𝗲𝘁𝘀. 𝗦𝗮𝗺𝗲 𝘀𝘁𝗮𝘁𝘀. 𝗪𝗶𝗹𝗱𝗹𝘆 𝗱𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝘁 𝘀𝘁𝗼𝗿𝗶𝗲𝘀. Visual inspection is 𝘯𝘰𝘵 optional. Anscombe's Quartet is a classic reminder of why plots matter: Each of the four datasets has: 👉The same mean for X and Y 👉The same variance for X and Y 👉The same correlation between X and Y 👉The same linear regression line But when you plot them? 🚨Completely different shapes: ✅A linear relationship ✅A clear curve ✅An outlier dominating the trend ✅A vertical line with a single influential point Same stats. Different stories. 𝗪𝗵𝘆 𝘁𝗵𝗶𝘀 𝗺𝗮𝘁𝘁𝗲𝗿𝘀 𝗶𝗻 𝘁𝗵𝗲 𝗿𝗲𝗮𝗹 𝘄𝗼𝗿𝗹𝗱: 👉KPIs may hide anomalies 👉Descriptive stats can misinterpret patterns 👉Decision-makers might rely on misleading summaries What looks like a tidy trend could actually be noise. Or worse: a trap. In data science, context is everything. And 𝘃𝗶𝘀𝘂𝗮𝗹𝗶𝘇𝗶𝗻𝗴 𝘆𝗼𝘂𝗿 𝗱𝗮𝘁𝗮 is often the fastest way to: ✅Spot errors ✅Identify outliers ✅Understand relationships Before trusting any model, always ask: 𝗛𝗮𝘃𝗲 𝘄𝗲 𝘀𝗲𝗲𝗻 𝘁𝗵𝗲 𝗱𝗮𝘁𝗮? 🎯 Plot first. Analyze second. Let's make this a norm: No summary statistics without visual context... ... especially in low-dimensional data. Curious to hear from others: Have you ever been fooled by stats that looked perfect on paper but broke down when you visualized them? Drop your favorite example below. #statistics #datascience #dataviz #analytics

So 10 years ago - my colleagues Scott Parrigon, Qiming Huang, James LeBreton and I proposed "Graphical Descriptives" to address data #transparency issues and the larger replication concerns in #psychology and #socialscience. The idea is this: #visualize your data - in a manner that protects #privacy - to reviewers and readers, not just your statistics. See here for our free, browser-based tool to make this easy for any researcher: 🔗 graphicaldescriptives.org 🔹 Upload your data. 🔹 Generate publication-ready and #OSF-ready visualizations. 🔹 No coding required. 🔹 Your data never leaves your browser.

A clean chart always beats a complicated one. Fancy visuals might look impressive, but they often hide the story instead of revealing it. Before adjusting colors, fonts, or effects, strip the chart down to its essentials. Start by asking yourself what’s truly needed for someone to understand the data fast. Here’s how to simplify effectively: - Remove extra gridlines and borders that don’t add meaning. - Keep only the most relevant labels and data points. - Sort data so the trend or comparison jumps out naturally. - Use consistent spacing and alignment to keep the layout steady. Save design tweaks for clarity, not decoration. When the clutter is gone, the insight shines through. Simplicity gives your audience less to look at and more to understand. Clean visuals aren’t plain, they’re powerful.

Satellite imagery captures reflected light across multiple spectral bands, with each band representing a data dimension. Sentinel-2 imagery records data in 13 spectral bands, each tailored to specific analytical purposes. While humans can visualize data in two or three dimensions (e.g., scatter plots or 3D plots), datasets with higher dimensions—such as Sentinel-2's 13 bands or hyperspectral imagery's 150+ bands—become challenging to interpret visually. Moreover, these datasets often contain redundant information and noise. Principal Components Analysis (PCA) is a mathematical technique that addresses these challenges by reducing the dimensionality of the data. It transforms the original spectral bands into a new set of uncorrelated components, known as principal components, which are ranked based on the variance they capture: Dimensionality Reduction: PCA condenses the dataset into a smaller number of principal components that retain most of the original variance, making analysis more manageable. Noise Suppression: By prioritizing the most informative components, PCA reduces the influence of noise and less significant variations in the data. Compact Representation: The transformed data is more compact, enabling easier visualization and improving computational efficiency for subsequent analyses. To demonstrate the power of PCA, I processed a Sentinel-2 image to create a PCA-transformed image. By assigning PCA1 to red, PCA2 to green, and PCA3 to blue, and applying histogram equalization for display, the resulting image reveals enhanced details that are invaluable for applications such as land cover classification, feature extraction, and change detection. Below is a comparison of the original true-color image and the histogram-equalized PCA image. Notice how the PCA image highlights subtle features, providing greater clarity and insight for remote sensing tasks #geospatial #gis #remotesensing #pca #dimensions

There's this assumption in bioinformatics that good EDA means exhaustive analysis. But here's the thing: the best exploratory data analysis isn't about doing more. It's about explaining less. The 3-slide test changes everything. Frame it as a 3-slide talk to a non-bioinformatician: Slide 1: What's in the dataset (samples, variables, source, structure) Slide 2: What patterns you see (clusters, gaps, batch effects, outliers) Slide 3: What actions to take (next steps, hypotheses, design flaws) For instance, when analyzing multi-omics data: Slide 1: "80 ovarian cancer samples, metastatic vs non-metastatic, with RNA-seq and DNA methylation data" (not technical pipeline details) Slide 2: "Found 1,200 differentially expressed genes, but only 180 overlap with methylation changes" (not exhaustive gene lists) Slide 3: "Focus on those 180 overlapping genes for biomarker validation" (not complex integration methods) This constraint forces you to simplify, clarify, and prioritize......fast. It cuts through analysis paralysis and gets straight to what matters. Because if you can't explain what you're seeing, you probably don't understand it yet. Try this on your next dataset. What story emerges when you strip away the complexity? #Bioinformatics #DataExploration #ScientificThinking #EDA #DataVisualization #CommunicationInScience #Omics #DataStorytelling #PrecisionMedicine #ComputationalBiology #ResearchTools

Choosing the right chart is half the battle in data storytelling. This one visual helped me go from “𝐖𝐡𝐢𝐜𝐡 𝐜𝐡𝐚𝐫𝐭 𝐝𝐨 𝐈 𝐮𝐬𝐞?” → “𝐆𝐨𝐭 𝐢𝐭 𝐢𝐧 10 𝐬𝐞𝐜𝐨𝐧𝐝𝐬.”👇 𝐇𝐞𝐫𝐞’𝐬 𝐚 𝐪𝐮𝐢𝐜𝐤 𝐛𝐫𝐞𝐚𝐤𝐝𝐨𝐰𝐧 𝐨𝐟 𝐡𝐨𝐰 𝐭𝐨 𝐜𝐡𝐨𝐨𝐬𝐞 𝐭𝐡𝐞 𝐫𝐢𝐠𝐡𝐭 𝐜𝐡𝐚𝐫𝐭 𝐛𝐚𝐬𝐞𝐝 𝐨𝐧 𝐲𝐨𝐮𝐫 𝐝𝐚𝐭𝐚: 🔹 𝐂𝐨𝐦𝐩𝐚𝐫𝐢𝐬𝐨𝐧? • Few categories → Bar Chart • Over time → Line Chart • Multivariate → Spider Chart • Non-cyclical → Vertical Bar Chart 🔹 𝐑𝐞𝐥𝐚𝐭𝐢𝐨𝐧𝐬𝐡𝐢𝐩? • 2 variables → Scatterplot • 3+ variables → Bubble Chart 🔹 𝐃𝐢𝐬𝐭𝐫𝐢𝐛𝐮𝐭𝐢𝐨𝐧? • Single variable → Histogram • Many points → Line Histogram • 2 variables → Violin Plot 🔹 𝐂𝐨𝐦𝐩𝐨𝐬𝐢𝐭𝐢𝐨𝐧? • Show part of a total → Pie Chart / Tree Map • Over time → Stacked Bar / Area Chart • Add/Subtract → Waterfall Chart 𝐐𝐮𝐢𝐜𝐤 𝐓𝐢𝐩𝐬: • Don’t overload charts; less is more. • Always label axes clearly. • Use color intentionally, not decoratively. • 𝐀𝐬𝐤: What insight should this chart unlock in 5 seconds or less? 𝐑𝐞𝐦𝐞𝐦𝐛𝐞𝐫: • Charts don’t just show data, they tell a story • In storytelling, clarity beats complexity • Don’t aim to impress with fancy visuals, aim to express the insight simply, that’s where the real impact is 💡 ♻️ Save it for later or share it with someone who might find it helpful! 𝐏.𝐒. I share job search tips and insights on data analytics & data science in my free newsletter. Join 14,000+ readers here → https://lnkd.in/dUfe4Ac6

Most researchers focus on writing. But the most persuasive part of your paper might not be in the words. I spend 30% of my paper preparation time on the core. The core of every paper is the results. It’s in the visuals, images, tables, graphs, and schemes. The results section isn’t just data. It’s your story engine. Here is a list of tools you can use to bring your story to life: ► Scientific illustration libraries & stock resources: The Noun Project: A Wide variety of icons, some suitable for simplified scientific representations. https://lnkd.in/eJyxwdh7 Bioicons: Specifically designed biological and medical icons. https://bioicons.com/ Freepik: Large collection of vectors and illustrations, including some scientific and medical content. https://www.freepik.com/ Simplify Sciences Publishing: Scientific illustrations and templates. https://lnkd.in/ebM5a4rg Servier Medical Art by Servier: Free, high-quality medical and biological illustrations. https://smart.servier.com/ ►Web-based tools (with illustration capabilities): Canva: A user-friendly design platform with vector elements and templates that are suitable for simpler scientific diagrams. https://www.canva.com Google Slides: Basic drawing tools for creating simple diagrams within presentations. https://lnkd.in/enPvsS6A Miro: A collaborative whiteboard platform with shapes and connectors is useful for creating conceptual diagrams and flowcharts. https://miro.com/ Biorender: A Specialized web-based tool with a large library of biological icons and templates for creating professional life science illustrations. https://www.biorender.com/ draw.io (now diagrams.net): Free, open-source diagramming tool for flowcharts and schematics. https://app.diagrams.net/ ► Installed software (advanced illustration): Adobe Illustrator: Industry-standard vector graphics software. https://lnkd.in/e9KY6KuE INKSCAPE: Free and open-source vector graphics editor, a powerful alternative to Adobe Illustrator. https://inkscape.org/ CorelDRAW Graphics Suite: Professional vector illustration suite (subscription and one-time purchase options). https://lnkd.in/ecPyAZmN ImageJ: Primarily for image processing and analysis in life sciences, but has basic annotation and drawing tools. https://imagej.net/ij/ Affinity Designer: Professional vector graphics software, a one-time purchase alternative to Adobe Illustrator. https://lnkd.in/epg2cDfh ► Specialized Installed Software: ChemDraw: For drawing chemical structures and pathways. https://lnkd.in/eqhhViW8 PyMOL: For 3D molecular visualization. https://www.pymol.org/ UCSF ChimeraX: Advanced molecular visualization. https://lnkd.in/eydbWgWF CellDesigner: For drawing biochemical networks and pathways. https://lnkd.in/e_QE9jsX ________ 📌 If what you need is proven strategy, support, and a community to grow in your academic journey, 𝗕𝗼𝗼𝗸 𝗮 𝗳𝗿𝗲𝗲 𝗰𝗮𝗹𝗹 𝘁𝗼 𝗮𝗽𝗽𝗹𝘆: https://lnkd.in/e-HnrCQW

Useful data science post Friday. I know you want to use cool visualizations. Whether you build dashboards, perform exploratory data analysis, or present evidence-based recommendations to the C-Suite via Powerpoint, the temptation to use complex visualizations is strong. This is simply because you work with data a lot, and your brain has acclimated to seeing data in 3-dimensional ways that can be difficult to visualize with bar graphs, line charts, and yes, even the much-maligned pie charts. Resist this temptation at all costs! Sankey diagrams, box plots, violin plots, radial charts, network graphs, tree maps, parallel coordinate plots and the like are impossible for most people to understand. Even dual-Axis line charts, waterfall charts, histograms, normalized (100%) stacked area charts and scatterplots should be used sparingly, as most end users will require a training and repeat exposure to understand them effectively. So, what should you use? What they already know and what's super intuitive. Single-axis line charts, vertical and horizontal bar graphs, and pie charts. Line charts should have a single axis and no more than five-to-seven lines, the fewer the better. Anything more than seven and most people will run into working memory issues. Less than five works for the majority of people. Bar graphs should not be stacked, but they should be grouped together. Use vertical bars for time series, and horizontal bars for comparing categories. When using horizontal bars, order them intuitively from greatest-to-least or vice versa. Pie charts should likewise have a limited amount of categories. Aim for a maximum of 5-6 with anything over that being labeled as "other." Because they are hard to read, ensure you have hover-over functionality that shows users the exact number on the pie chart (both absolute and as a percentage). That will help avoid misinterpretations. So, it's not sexy advice, but it's the best advice for delivering usable content to end users. Enjoy your day! #datascience #datavisualization #powerbi #tableau #qlik

Before we illustrate, we often ask uncomfortable questions: - What can we remove? - What’s doing too much work here? - Is this figure trying to impress or to explain? In scientific visuals, clarity rarely comes from adding more. It usually comes from subtraction. Every extra label, panel, color, or effect competes for attention. And when everything is important, nothing is. The most effective figures don’t show everything you did. They guide the reader to what actually matters. That’s where real communication happens! #ScientificIllustration #DataVisualization #ScienceCommunication #FigureDesign #ResearchDesign #DrawImpacts #Biotech #LifeSciences #ClarityOverComplexity Figure illustration by our studio DrawImpacts for Dr. Pashtoon Kasi