Along with about 35 million other Americans, I’ve been watching the Olympics in primetime every night after work.
And while NBC has done a notoriously bad job at keeping the production fair and balanced, it’s still pretty easy to tell that some countries are pulling in way more medals than others.
Currently, the United States has the most gold medals while China has the most overall medals and Great Britain is third in both categories. But these totals are changing every few hours, so check out the latest numbers here.
Seeing the same countries on the podium made me curious: are there patterns between the countries that bring home the most medals?
Specifically, if you controlled for the size of a country’s economy, would you be able to predict the number of medals they’d win?
Could Olympic prowess be a leading indicator of a country’s financial success? Or maybe it worked the other way, and a growing economy was a sign of increasing athletic competitiveness?
The questions were piling up and I needed to find answers.
I knew this wasn’t going to be as rigorous as an academic paper — it was just scratching an itch for me — so I started my data collection on Wikipedia.
I build a python scraper that pulled data from Wikipedia’s “Summer Olympics medal table” for every year back to 1980. I collected both the total number of medals, as well as the number of gold medals for each country in each year.
I decided to skip Winter Olympics since many countries are at a decided disadvantage because of their climate. The games of the Summer Olympics didn’t seem to favor any particular regions.
Once I had that data, I needed a way to control for the size of a country’s economy. I needed time-series data, but I couldn’t find it on Wikipedia, so I turned to the World Bank instead. They had exactly what I needed: Real GDP for every country and year back to 1980.
It took a bit of work to re-arrange the data in a way that I could plot it for each country over time, but once I had it all laid out in excel, I started making graphs.
The raw numbers were hard to draw meaningful conclusions from — imagine a graph whose Y-axis has numbers in the teens as well as the tens of billions — so I converted everything to percentages and tried again.
For countries that occasionally won a medal or two, there was too much variance to really notice anything meaningful.
In an Olympics games, about 900 medals are awarded (300 of which are gold), so a country who picked up six medals one year and only three the next would see their total medal percent drop from 0.67% to 0.33%. However, most GDPs for countries in this range were way more consistent, fluctuating by only tenths of a percent.
But when you start looking at some of the bigger countries with more consistent medal winnings, some patterns emerge.
I originally imagined the percentage of gold and total medals would trend up or down for a given country over time, however it was fairly consistent, even across decades.
Unfortunately, this meant that it was hard to detect any interaction between GDP and Olympic dominance — with the usual caveat that correlation wouldn’t even imply causation anyways.
However, what I did notice was that some countries continually over-performed at the Olympics, while others continually under-performed.
By graphing the percentage of medals a country wins against their percentage of world GDP, it was easy to see which countries were more dominant athletically than they were economically, and vice versa.
Surprisingly, the United States has actually been under-performing at the Olympics, bringing home only 10%-15% of the medals, compared to our 25% of the world’s GDP.
China has been performing just about as expected, bringing home more and more medals each year as their economy grows.
Other countries to note were Australia, who has grown increasingly dominant at the Olympics while their economy has remained largely stagnant.
Also Kenya, who have consistently brought home about 1% of the medals despite having a teeny tiny economy.
At this point, there’s still a lot than can be done with this data. For example, what if you also looked at trends in population, or land mass? Would larger countries bring home more medals, or is GDP more closely related to Olympic performance than size?
It would be cool to find some metric that seemed to trend closely with Olympic medals winnings. Then you could potentially make predictions about that metric based on how a country performs every 4 years, or conversely, predict a country’s medal winnings based on that metric.
But as you can imagine, the athleticism, determination and skill required for each Olympian to bring home a medal is largely determined by their own personal circumstances, and not necessarily those of their country in aggregate.
What do you think, does this surprise you?