Obesity — a Data Story
Rescuing nuanced pattterns from the clutches of a bad graphic
“This figure may give you a migrane”
Sometimes, I get emails that look like this
Sent: Monday, July 29, 2019 at 07:59 From: Jasleen Grewal Subject: This figure may give you a migrane As you can see, 100% of the graphs are ineffective.
Here, I wanted to take you through my reaction to the figure, which was quick, and the redesign, which wasn't quick.
text labels — it's a hard life
I'm always on the lookout for abused text. So here I cried. A lot.
strangely structured legend
Do we really need a footnote inside the legend? The globe? The hyphenated "Body-Mass-Index". By this point, I really could feel that migrane.
here's the graphic — now what?
What question's does this figure answer? Here's my list, with answers.
1. How many countries are there in the world? A lot.
2. What is the range of BMI ≥ 25 prevalence? 18—89.
3. Who has the lowest and highest prevalence? Vietnam and Nauru.
4. What is the median prevalence? Probably 55 and answering this is only made easy by the fact that the book's spine splits the plot into largely two equal halves
5. What is the prevalence where I live (e.g. Canada)? I gave up trying to find "Kanada".
Essentially, the two-page figure of ring charts is equivalent to the summary
critique by redesign
It's obvious what's wrong with the figure. How do you fix it?
Using the list of countries by body mass index, I created a poster that tells interesting stories about how high BMI and obesity vary across countries and genders.
I describe the design and stories in the poster in the design section.
Symmetric alternatives to the ordinary least squares regression
What immortal hand or eye, could frame thy fearful symmetry? — William Blake, "The Tyger"
This month, we look at symmetric regression, which, unlike simple linear regression, it is reversible — remaining unaltered when the variables are swapped.
Simple linear regression can summarize the linear relationship between two variables `X` and `Y` — for example, when `Y` is considered the response (dependent) and `X` the predictor (independent) variable.
However, there are times when we are not interested (or able) to distinguish between dependent and independent variables — either because they have the same importance or the same role. This is where symmetric regression can help.
Luca Greco, George Luta, Martin Krzywinski & Naomi Altman (2025) Points of significance: Symmetric alternatives to the ordinary least squares regression. Nat. Methods 22:1610–1612.
Beyond Belief Campaign BRCA Art
Fuelled by philanthropy, findings into the workings of BRCA1 and BRCA2 genes have led to groundbreaking research and lifesaving innovations to care for families facing cancer.
This set of 100 one-of-a-kind prints explore the structure of these genes. Each artwork is unique — if you put them all together, you get the full sequence of the BRCA1 and BRCA2 proteins.
Propensity score weighting
The needs of the many outweigh the needs of the few. —Mr. Spock (Star Trek II)
This month, we explore a related and powerful technique to address bias: propensity score weighting (PSW), which applies weights to each subject instead of matching (or discarding) them.
Kurz, C.F., Krzywinski, M. & Altman, N. (2025) Points of significance: Propensity score weighting. Nat. Methods 22:638–640.
Happy 2025 π Day—
TTCAGT: a sequence of digits
Celebrate π Day (March 14th) and sequence digits like its 1999. Let's call some peaks.
Crafting 10 Years of Statistics Explanations: Points of Significance
I don’t have good luck in the match points. —Rafael Nadal, Spanish tennis player
Points of Significance is an ongoing series of short articles about statistics in Nature Methods that started in 2013. Its aim is to provide clear explanations of essential concepts in statistics for a nonspecialist audience. The articles favor heuristic explanations and make extensive use of simulated examples and graphical explanations, while maintaining mathematical rigor.
Topics range from basic, but often misunderstood, such as uncertainty and P-values, to relatively advanced, but often neglected, such as the error-in-variables problem and the curse of dimensionality. More recent articles have focused on timely topics such as modeling of epidemics, machine learning, and neural networks.
In this article, we discuss the evolution of topics and details behind some of the story arcs, our approach to crafting statistical explanations and narratives, and our use of figures and numerical simulations as props for building understanding.
Altman, N. & Krzywinski, M. (2025) Crafting 10 Years of Statistics Explanations: Points of Significance. Annual Review of Statistics and Its Application 12:69–87.