Hello and welcome to another edition of The Crunch.
In this week’s newsletter, we have charts on the countries most exposed to Trump’s tariff push, how much it costs for a band to go on tour, the staggering increase in solar electricity generation capacity around the world, the extent to which firings of US government officials overlaps with departments investigating Elon Musk’s business interests, and how to turn any chart into a skate park.
But first … who is leading in the polls?
We have just finished a big redesign of our tracking page for Australian federal election polls. The redesign aims to make the charts clearer, de-emphasise the average (mean) and highlight the uncertainty in polls and our modelling.
You can see the live version here, but we’ll go through the key aspects of the redesign.
The biggest change is we removed (Nick’s favourite) diamond chart. This chart previously showed the mean, the credibility interval (sometimes also known as a margin of error) and compared the current polling estimate with each party’s result at the past election:
While we’re still fans of this design – we used it in some of our Covid tracking pages – it falls over if you have to take away the diamonds, which we wanted to do to reduce the emphasis on the mean.
We’ve always wanted to communicate to readers that polls contain some uncertainty – and there’s still uncertainty even in our poll of polls. When we highlight the mean, even if an error bar or credibility interval is included, the mean will be the only takeaway for many.
So it is now two charts! The first new chart shows only the range (credibility interval) of possible values for the estimated vote, and (possibly controversially) doesn’t show the mean at all:
The second chart now shows the difference between the current estimated vote, and each party’s result at the last election, essentially showing the “swing” to and away from each group:
While we do have a mean, and “error bars”, we’ve hopefully over-emphasised the credibility interval to such a degree that people can’t possibly miss it.
Some of this was prompted by John Burn-Murdoch’s column about uncertainty in polling ($). It’s something we were already trying to emphasise, but we decided that wasn’t going far enough – please let us know what you think.
Four charts from the fortnight
***
1. The who and what of global trade
Our colleagues in the UK put together this series of visuals to explain how the US trade deficit with some countries has changed and what the impact could be. Most of the focus is on Canada and Mexico but Australia does make a Where’s Wally appearance in this nice bubble chart:
***
2. Make Touring Great Again
We’ve been hearing for years that there’s little money for musicians in streaming, and the real money is in live events. Well, it turns out there’s a huge amount of overhead in touring, too. Shaylee Safie has broken this all down on Datawrapper. Even with some dire assumptions – such as the band all couch surfing on tour – the numbers don’t look great for a lot of smaller bands.
By the way, if you aren’t reading the Datawrapper blog, you really should.
***
3. Who watches the watchers?
There’s been a lot of chaos in the less than a month since Trump returned to power. Not least at the more than 11 federal agencies that have been investigating Elon Musk’s companies or that have pending complaints or lawsuits. Here’s a handy flow chart from the New York Times so you can keep track.
***
4. Prices Install times are down down
A lot of climate visuals show the world going in the wrong direction – polluting more and warming. But this chart from Our World in Data, thankfully, doesn’t. In 2004 it took about a year to install a gigawatt of solar power but by 2023 this was only taking a day on average.
Max Roser published an explainer on the growth in renewables a few years ago that is still a good read. And there’s a bunch of other climate and energy data to explore at Our World in Data if you are so inclined.
Our bookmarks from this week
Another cool project turning charts into physical objects – beaded bracelets
A deep dive into whether people actually do hate Coldplay
Data visualisation is an art
A handy guide that unpacks some of the decisions, abstractions, limitations and assumptions people often make when building charts
Off the Charts
This started as a joke on Bluesky but it’s now a real thing! SK8plotlib is a Python library that will turn any matplotlib line chart into a line skateboarding game, built by Emily Hunt:
Still trying to figure out if this makes it any easier to format decent-looking dates in matplotlib. We might be sold, if so.
Sign up
If you would like to receive The Crunch to your email inbox every fortnight, sign up here.
