As 2020 draws to a close, we are reflecting on a lot of things at Riffyn — and I don’t mean pandemics or elections (although we have all, indeed, reflected deeply on those things). The new year, perhaps more than ever, brings the promise of freshness and newness, and we’ve been thinking about how to bring fresh, new, and relevant content to you.
Since we love data, we dove into our content analytics to learn which topics that you love to read about — and those you don’t.
To find the 5 most popular Riffyn blog posts of all time, I asked Google Analytics to go as far back as October 2017, when our CEO Tim Gardner wrote Riffyn’s very first blog post, “Buried in Data and Starving for Information.”
Below are the top 5 most read Riffyn articles of all times. Two articles consistently outperformed any other post — and continue to receive more reads than our other posts on a monthly basis. One focuses on false discovery rate, the other on family-wise error rate.
The continued popularity of these articles, despite being three years old, taught us that you LOVE educational posts that demystify complex statistical/data analysis concepts. We’re taking this finding seriously, and are committing to bringing you more similar content in 2021.
Until then, I welcome you to take a walk down memory lane and visit (or revisit) our most popular articles of all time!
“FDR is a very simple concept.” So begins our most popular blog article of all time. Clearly, many came to the article with the perception that FDR — false discovery rate — was anything but a simple concept. We hope that changed by the time our readers had finished!
In the article, our CEO Tim Gardner delves into the nitty-gritty details of FDR — what it is, how to use it, what each piece of the equation means, and how to use and calculate Q values. He even includes some code that you can use to calculate FDR using Excel. Perhaps one of the most actionable, value-packed blog posts ever written. Check it out for yourself if you haven’t already!
“Family-wise error rate.” Four words that, unless you are a statistician for a living, are probably just as daunting as “false discovery rate.” But much like he demystified false discovery, Riffyn CEO Tim Gardner similarly demystified family-wise error rate in our second most popular blog article of all time.
In this post, Gardner explains what family-wise error rate is and does, discusses one of the most popular family-wise error rate methods (Bonferroni correction) and why you might just want to reconsider using false discovery rate for your experiments instead. Read (or re-read) the article to find out why!
False discovery rate made an appearance again with our third most popular article of all time. In this article, CEO Tim Gardner uses a hypothetical situation based on an actual experience of one of our customers to illustrate the likelihood of false positives and false negatives in experimentation and how to control for them.
Gardner discusses the Benjamini-Hochberg method of FDR calculation, compares and contrasts p- and q-values, and highlights the importance of experimental design when attempting to minimize both false negatives and false discoveries. This is a great article to review time and again … starting right now!
Riffyn was founded on the desire to make science better: faster, more reproducible, and more efficient. But, as CEO Tim Gardner wrote on October 3, 2018,
“.. there is one problem. Until today, Riffyn SDE was available only to paid subscribers. It is quite obvious to us that this closed posture of Riffyn SDE is incompatible with our mission to reshape and improve modern scientific practices.”
To that end, Riffyn made its Process Data System, Riffyn Nexus (then known as Riffyn SDE) available free-of-charge to any person affiliated with a non-profit organization. Riffyn Nexus may have been developed with industrial R&D in mind, but the fact that this article is our fourth most popular of all time shows that the Riffyn principles are those that all scientists can get behind. Read the article to learn about all of those principles — and how theys are enabling scientists around the world to do better science.
It is fitting that our fifth most-read blog post of all time is our very first post — CEO Tim Gardner’s heart-to-heart with the world about the state of how modern science is done, and how and why it can and should be done better. Comparing the process by which scientists perform, record, and share their work to the way computer scientists and engineers do, Gardner makes a call to all scientists to rethink the scientific experiment.
What would you like to see us write about in 2021? Send your suggestions to firstname.lastname@example.org — we'd love to hear from you!
Embriette is an academic-turned science writer with a passion for spreading responsible science. She holds a PhD in microbiome research from Baylor College of Medicine. After a 4-year post-doc, during which she managed the world's largest citizen science research project (the American Gut Project), Embriette became a full-time science writer and research consultant. You can find her work at riffyn.com, synbiobeta.com, and her personal webpage: drhydenotjekyll.com