Olympics and Olympics' Hosts

Sep 11, 2021

Does hosting the Olympics improve performance in the medals table?

Tokyo 2021

Before delving in to my investigation, let’s take a look at how the medals table finished at Tokyo 2020 (2021).

We see familiar nations at the top: The USA, China, the United Kingdom, and notably, Japan.

Let’s take a more detailed view at the medal tables, breaking down the medals into bronze, silver, and gold.

Ceeating a new dataframe containing host nation data

Hosting The Olympics

Hosting the olympics is an incredible honor for a country to have.

The previous three Olympics were hosted by: Japan, Brazil, and the United Kingdom.

Here’s an overview of their Olympic histories, with the years they hosted highlighted

Overall: Summer Olympics

Now I’ll widen my view.

Here’s how the leaderboard stacks up when we consider all Summer Olympic games.

Let us now view the distribution of medals on a map.

The darker the color, the more medals a country has won

Next, I want an overview of how each country performs; when they host the olympics, when they don’t, and on average.

First, let’s look at all nations over time

What does this show us?

We see that host nations seem to attain higher medal totals than other nations.

However, we must view this with caution, as it’s often the wealthy nations that host the games, therefore they are expected to do better anyway.

Host Nations

So we know that several nations seem to consistently perform well at the Olympic games. But that’s not what this project is seeking to confirm.

It’s now time to shift focus to those who have hosted the games.

What does this show us?

This is a fantastic plot. I’ve focussed only on nations that have hosted the olympic games.

We gain so much information from just one visual. So, what do we see:

  • Historic performance (grey dots)

  • Average performance (blue dots)

  • Performance when hosting (red dots)

It seems clear that when a nation hosts the olympics, they perform better!

Let’s view the host countries geographically

Does performance vary when hosting & not hosting?

Let’s take a sample from the new dataframe and plot the average medal totals for each nation, both when they’ve not hosted the olympics, and when they have.

What does this show us?

This visual is very interesting.

We see that all but one nations drastically imrproved their average total medal tally when they hosted an olympic games.

Canada appear to be an anomaly - they performed much worse when they hosted.

Medal Performance Over Time for Host Nations

We’ve looked at static averages, but now let’s view the games as they happened.

First, we can view the USA’s medal tally over time. I’ll highlight the games that they hosted.

What does this show us?

Depsite consistently performing well, it does look as though the USA performs even better when they are the tournament hosts.

What about other nations. Let’s try the UK

What does this show us?

Again, it appears that hosting does help performance.

However, the UK was on an upward trend during the 2000’s, so perhaps this would have continued anyway

What about Gold medals?

I’ve focussed on Total Medals so far, as I beleive all medals are impressive.

However, we can view each separately…

What does this show us?

The UK, when hosting in 2012, had a great performance, with the highest number of gold medals ever. This was then surpassed in Rio 2016 though.


Let’s move away from totals, and towards proportions

What does this show us?

As mentioned previously, hosting does appear to have a positive effect on total medals, but this looks to be true of gold medals too.

What about Japan?

As I mentioned at the beginning, I was impressed by Japan at this olympics. Did they perform better than usual?

Interestingly, Japan’s total medal count is not too extraordinary compared to previous olympics. Though it is still impressive, in Japan’s last 5 olympic appearances it would rank 4th.

What about gold medals? This is, afterall, how the medal table it often calculated

What does this show us?

Japan, hosts for 2020 (2021), had a fantastic games in terms of gold medals; winning the highest proportion of golds for over 40 years.

What does this show us?

Not only is the proportion of gold medals higher than usual, so too is the absolute number. This was also the case when Japan hosted the games back in the 60’s, too. We can view this more clearly below

Statistical Testing

We have seen through our visual journey that Hosting thw Olympics does seem to lead to an increased performance, but we will now do some statistical tests to find out whether or not these differences could be considered significant or not.

What does this show us?

This shows what we’ve seen throughout this analysis.

Using only countries who have hosted an olympic games, what is their average medal tally when Hosting & Not Hosting. It is clear that Hosting provides the higher tally, but contries host very infrequently, therefore, we must conduct statistical tests to be sure of this intuition

What does this show us?

We note that the medals are not distributed normally. They are positively skewed. To mitigate against this, I will take the natural log of the total medals.

Key Statistics Mean Natural Log of Total Medals when Not Hosting: 3.086281690114543 Mean Natural Log of Total Medals when Hosting: 4.566512430214569 Standard Error of Natural Log of Total Medals when Not Hosting: 0.06927049040753175 Standard Error of Natural Log of Total Medals when Not Hosting: 0.19153671012884035

Is this difference enough to declare that the two means are from different distributions?

Standard Errors between means 7.26750565554929

This means that it is incredibly unlikely that the two means are from the same populations.