Science

NOVEL DIGITAL BIOMARKER FOR UNDERSTANDING MENSTRUAL CYLCES I’m excited to share groundbreaking WHOOP research published in Nature’s Digital Medicine journal earlier this week. This work is a big leap forward in how we understand and monitor female physiology. It was previously well understood that several vital signs, (most notably temperature, but also importantly Heart Rate and Heart Rate Variability) fluctuate with the phase of your menstrual cycle, but it was never before understood whether or not the extent of those fluctuations were significant and if so, what they might mean. We set up a study with WHOOP data harnessing the insights from 11,500 women who opted into our research program. By analyzing over 45,000 menstrual cycles, our team was the first to describe the significance in the amplitude of those fluctuations, and in doing so, develop a novel metric that may change the role wearables play in female reproductive health. Amplitude was observed to be suppressed in individuals with characteristics of reduced fertility, such as higher BMI and older age. This completely non-invasive marker could one day be used to identify irregularities in reproductive health earlier, reducing time to diagnosis. This study was made possible by our always-on data and global scale. It utilized over 1 million days of WHOOP data, enabling insights that aren’t feasible without a 24/7 wearable. Only 3% of medical research is focused on women. Whoop is committed to investing equally in women’s and men’s health research. We’ll continue to share women’s research discoveries like this and more. We use this research to inform and guide changes to our product. Thank you to our incredible research team, led by Emily Capodilupo, for their dedication and innovation. Check out the full study here: https://lnkd.in/eXSqC4-5 #WHOOP #wearabletech #research #WomensHealth #innovation

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

AI Fingerprints Found in Millions of Scientific Papers, Study Reveals Introduction: The Quiet Rise of AI Authorship in Academia As large language models (LLMs) like ChatGPT and Google Gemini become increasingly capable of producing high-quality writing, their influence is now visibly permeating academic literature. A massive study analyzing over 15 million scientific papers has uncovered measurable linguistic patterns that suggest a significant portion of biomedical research may already be shaped—at least in part—by artificial intelligence. Key Findings from the Study • The Scope of the Analysis • Conducted by U.S. and German researchers, the study focused on biomedical abstracts published in PubMed—one of the largest databases of peer-reviewed life sciences literature. • Researchers used AI-detection techniques to track stylistic patterns and word choices consistent with LLM-generated or LLM-assisted writing. • AI’s Growing Influence in Scientific Writing • The results show that at least 13.5% of scientific papers published in 2024 were likely produced with help from a large language model. • Since the emergence of tools like ChatGPT, there has been a marked increase in specific phrasing and terminology that mirrors LLM outputs. • These “AI fingerprints” suggest that AI-generated or AI-edited content is becoming increasingly normalized in academic publishing. • Implications for Research Integrity and Peer Review • The widespread use of LLMs in academia raises questions about authorship transparency, originality, and peer review standards. • Editors and journals may need to establish disclosure protocols and develop more robust AI-detection tools to maintain the integrity of published work. • While AI can assist with grammar and structure, overreliance could blur the line between assistance and authorship. Conclusion: Rethinking Scientific Writing in the AI Age This study provides compelling evidence that AI is no longer a background tool in academic research—it’s rapidly becoming a co-author. As scientific publishing adapts to this new reality, the challenge will be to harness AI’s efficiency without compromising intellectual integrity, accountability, or the human creativity that drives true discovery. https://lnkd.in/gEmHdXZy

CRA Tips for SIV preparation! SIV preparation is a comprehensive, methodical process to ensure the most effective SIV conduct. These tips will help facilitate the process. 👉1.     Read and confirm the exact process for SIV conduct from the monitoring plan. The monitoring plan should detail all visit aspects: systems utilized (EDC, ePRO, lab portals, central imaging vendors), protocol elements for review, investigator site file instructions, IP receipt, temperature monitoring, inventory, and confirmation instructions. 👉2.     Stay apprised of site activation status. 👉3.     Ensure the PI, and study coordinator/site staff are aware of any ancillary site staff that need to attend the SIV (pharmacy staff, raters, sub investigators). 🔥4.     Prepare an agenda detailing topics/categories/tasks for review/completion, time periods for each activity and whom from the site is required to attend. Organize timing and planning with the study coordinator so they can plan accordingly. Construct an adaptable agenda as unexpected changes may occur, and you will need to accommodate. Flexibility is key! 👉5.     Send required documents for site completion in advance of the visit. This will save everyone time and ensure at least some can be collected at the visit. 🔥6.     Track supply shipments and site receipt (lab kits, investigator site file, ePRO devices, IP, etc.), for inventory at the SIV. Check with the site before the visit regarding receipt or notification. 👉7.     Check site staff systems access to EDC, IVRS, SIP-whatever platforms are being used. Be sure to provide help desk information. 🔥8.     Prepare an abbreviated PI SIV presentation-just in case. Mark the slides and information required for PI review. Unexpected things happen and you may have a shorter amount of time with the PI to review protocol/study information. You can review the remaining information with site staff during the visit. 👉9.     If a sponsor representative is attending, be sure to liaise with them and provide them with the agenda, directions to the site and relevant information. 🔥10.  Bring printed copies of slides/material and ensure you have sent the slides and protocol to the site before the visit. 🔥11.  Arrive to the SIV early enough to allow for AV set up. The site may have a large monitor/screen that they want you to use for the protocol presentation to accommodate a large audience. If for some reason your laptop cannot connect successfully with their equipment, ensure you have brought a secondary monitor to compensate for this viewing. 

In #datavisualization, the bar chart 📊 is one of the most popular chart types. It’s a staple in dashboards and reports and is often used to make precise comparisons. Bar charts are equally prevalent in #datastorytelling. However, assessing how we can reduce our audience's cognitive load is crucial—even with something as familiar as a bar chart. The accompanying example illustrates a common scenario: comparing two data series side-by-side in a grouped bar chart. Typically, this involves comparing actual values against forecasted, targeted, or budgeted figures. With the grouped bar chart, both actual sales and forecasts are clearly displayed, requiring the audience to simply compare the heights of the two bars. It's not too difficult to interpret, right? But with the bar chart with markers, the forecasts are overlaid on top of the actual sales. This streamlined approach reduces cognitive load, making it quicker and easier for your audience to identify discrepancies between actual performance and forecasts. While the effect may seem subtle, it can make a significant difference when multiplied across multiple charts in a data story. I’d rather my audience focus their mental effort on understanding key insights than deciphering charts. Simplifying comparisons is key to making visual information more accessible and impactful, especially in data stories. How often do you consider the cognitive load of your data visualizations? What other small changes do you use to reduce cognitive load? 🔽 🔽 🔽 🔽 🔽 Craving more of my data storytelling, analytics, and data culture content? Sign up for my newsletter today: https://lnkd.in/gRNMYJQ7

✍️ Submitting to a top CS conference or journal? I just released an open-source, concrete, and opinionated checklist for CS paper writing — designed to prevent desk rejects, improve clarity, and save co-authors and reviewers a lot of pain. I have asked all my lab papers to go through this checklist before submission. 🧠 Inspired by real (painful) examples: 1. Forgot to include a co-author before submission (happened much more frequently than you can imagine) 😬 2. Revealed author identity via GitHub repo metadata 🔍 3. "Novel framework..." + no baselines + one giant equation = 🚫 4. Copy-pasted LLM citation hallucinations that don’t exist 🧨 ✅ The checklist covers many things (if not everything): title, abstract, method, experiments, figures, references, hallucinated citations, and final sanity checks. 📄 English & Chinese versions available. 🌐 GitHub: https://lnkd.in/gaQ85ChY Use it. Share it. Improve it. Save a paper (or a career). #CSResearch #AcademicWriting #MachineLearning #PhDLife #PeerReview #LLM #Reproducibility #OpenScience

I'm a recovering "over-thinker" of every post.  I'd stare at the screen.  Delete sentences.  Rewrite them.  Then delete them again.  "I suck at this."   "People won’t like this."   "This sounds so dumb."  So I’d scrap the post.  Promise myself I’d try again tomorrow.   Then repeat the same cycle the next day. And of course delay the fix. It was painful. Eventually, I forced myself to hit publish.     Small engagement.   Little momentum.   Tiny traction.  So I overthought even more.    Maybe I needed better hooks.   Maybe I wasn’t being persuasive.   Maybe I just wasn’t cut out for this.  Then, I tried something different.  I stopped trying to sound smart.   Stopped overanalyzing every word.   And quit worrying about "perfection." Instead, I told simple stories.    Niche relatable stories.  With simple and relatable lessons. Ones that tied back to my experience.  These stories were not only engaging.   They were relevant.  And built real rapport with my niche.  That’s when things changed.  No fancy copywriting.  No marketing tricks.   No viral tactics.  Just clear, honest, simple, real, human...  Storytelling.  Do this with 3 steps: 1) Start with a real moment Something specific that actually happened 2) Show the struggle (not just victory) Add frustrations, difficulties, & uncertainties 3) End with the insight (don't miss this) Get this right by adding insights + actionable tips Tell your audience exactly what they need to hear. Do it, & you’ll never run out of content again.

I used to think colorful charts helped display information. Now I know they do exactly the opposite. When it comes to data visualization, color *is* crucial. But not in the way you’ve likely been taught. The general rule of thumb is that you should use color sparingly and strategically. In other words, never use color for the sake of being color*ful*. Here’s how: First, identify your core colors (I recommend 1-2 max): Option 1 ↳ Use your company’s (or client’s) brand colors. This is often the easiest and best choice. (But remember, you don’t have to use *all* the brand colors.) Option 2 ↳ Use an online color palette (check out the resources linked in the comments to get started). I’ve also searched Pinterest for things like “blue and green color palettes.” Second, follow best practices: Use grey as your default.  ↳ Create all your charts in greyscale first. Then, incorporate color to draw your audience’s eyes to the most important takeaways or data points. Use 1-2 core colors throughout your presentation.  ↳ Use your core colors to highlight the specific trends, categories, or insights you want your audience to pay attention to. Be aware of cultural associations.  ↳ Color symbolism varies across the globe - for example, red often carries a negative connotation in Western cultures, but represents luck and prosperity in Eastern/Asian cultures. Be mindful of color blindness.  ↳ Approximately 8% of men and 0.5% of women are colorblind (red-green being the most common). In general, less is more. Imagine someone were to look at your chart and say “Why is THAT particular bar blue? Why is THAT one green?” If you can’t give a clear answer, it's time to go back to the drawing board. —-— 👋🏼 I’m Morgan. I share my favorite data viz and data storytelling tips to help other analysts (and academics) better communicate their work.

On kindness in peer review: 9 better ways to say “This paper needs work" Every so often, I come across a reviewer comment that calls a contribution trivial or says it “does not rise to the level expected” at a journal. When I see that language, I wince. Even if the critique has merit, it often overshadows otherwise valuable points in the review. Why? Because it makes the authors feel like the entire review team—not just one reviewer—didn’t see any merit in their work. So, what can we do instead? To help authors actually use your feedback? Soften your tone—not your standards. Use language that clearly signals concern about the contribution without shutting down the possibility for improvement. Rather than making the author angry, use language that engages the author with your comments and encourages them to improve their work. Here are nine thoughtful phrases I’ve seen good reviewers use this past year, that encourage engagement. They’re especially useful in peer review, mentorship, or conference feedback: 1. "The core argument feels underdeveloped, and I had trouble fully engaging with it." This gently signals the paper didn’t land, while pointing to a fixable issue. 2. "I struggled to connect with the contribution—perhaps more framing or positioning could clarify its relevance." Invites the author to sharpen the positioning of their work. 3. "The paper raises important questions, but the current structure makes it difficult to appreciate its full impact." Encourages authors to revise the structure for better clarity. 4. "I found myself wanting more clarity on how this piece fits into the broader conversation." Suggests adding context. Consider: “It doesn’t resonate with me because the context is missing.” 5. "This may reflect my own disciplinary perspective, but I had difficulty connecting with the theoretical framing." Acknowledges your own lens and invites the author to strengthen their framing for a wider audience. 6. "The writing is thoughtful, but I had trouble seeing how the pieces come together to form a cohesive narrative." Encourages a shift from listing elements to telling a coherent story. 7. "The manuscript feels preliminary—there’s potential here, but it’s not fully realized yet." Flags underdevelopment without sounding dismissive or harsh. 8. "The contribution may benefit from more grounding in empirical or theoretical detail to fully resonate with readers." Only use this if you can specify what detail is needed. 9. "This version didn’t quite land for me, but I believe with revision and sharper focus, it could really shine." Provides an honest, hopeful invitation to revise. Never forget. Reviewing is about stewardship. It’s about helping authors make their work stronger—even when it’s not there yet. So rather than tearing down papers, offer a well-phrased critique, that encourages authors to keep working. #PeerReview #AcademicWriting #AcademicJourney #AcademicCulture

The silent productivity killer you've never heard of... Attention Residue (and 4 strategies to fight back): The concept of "attention residue" says there is a cognitive cost to shifting your attention from one task to another. When our attention is shifted, there is a "residue" that remains in the brain and impairs our cognitive performance on the new task. Put differently, you may think your attention has fully shifted to the next task, but your brain has a lag—it thinks otherwise! With apologies to any self-proclaimed proficient multitaskers, the research is very clear: Every single time you call upon your brain to move away from one task and toward another, you are hurting its performance—your work quality and efficiency suffer. Here are four strategies to manage attention residue and fight back: 1. Create a Boot Up Sequence Your personal boot up sequence is a series of actions that prime your mind and body for deep focus work. For me, this involves cold brew coffee, classical music, and sitting in a bright, well-lit environment. Create your own boot up sequence and your attention performance will improve. 2. Focus Work Blocks Block time on your calendar for sprints of focused energy. Set a timer for a 45-90 minute window, close everything except the task at hand, and focus on one thing. It works wonders. 3. Take a Breather Whenever possible, create open windows of 5-15 minutes between higher value tasks. Schedule 25-minute calls. Block those windows on your calendar. During them, take a walk or close your eyes and breathe. 4. Batch Processing You still have to reply to messages and emails. Pick a few windows during the day when you will deeply focus on the task of processing and replying to these. Your response quality will go up from this batching, and they won't bleed into the rest of your day. Attention residue is a silent killer of your work quality and efficiency. Understanding it—and taking the steps to fight back—will have an immediate positive impact on your work and life. 📌 To learn more science-backed systems to improve your life, join thousands who have preordered my first book: https://lnkd.in/eGhQwaRC Enjoy this? ♻ Repost to help your network and follow Sahil Bloom for more.