Data storytelling tips from Election Charts
Hello and welcome to the 83rd Sweet Spot.
It's the UK General Election on Thursday. We've been bombarded with data graphics from the media and the political parties: what can we learn from the best (or worst) examples? Read on...?
But first, some notices:
The image on the left is from Savanta. When lines are thin, the colours for Reform (teal) and Conservative (blue) are hard to distinguish. It's not helped by the legend also being very small. The solution? Thicker lines, label the lines, and colour-encode the title. I explain more in this YouTube short.
Donuts have their place. But at least make the labels readable: the example from Survation (left) above could have been much better if the labels were bigger (my makeover, right). Check out my video outlining a different donut challenge from a Savanta chart.
Stacked bars are excellent but you, the author, need to carefully curate the message by choosing the order of the stack sections and colour choice. Above are examples reworking a Survation chart. I explain more in this video.
How about chart crimes? One egregious chart came from JLP, who conducted polls for The Rest Is Politics. This example has three serious problems. Check out my my makeover ideas on LinkedIn.?
Coming up:
Eli Holder is our next Chart Chat guest. He's an expert on misleading visualisation and an amazing speaker. We'll be doing live psychology experiments to find our own biases!?Click here to register. We're live on July 18 at 11am EST.
Crafting a data story is about how to apply different LEVERS to your charts. Check out my recent Tableau webinar.
I've been taking Claude and ChatGPT for a ride recently. How good are they getting at data analysis?
Software Engineer | Python Developer | Cloud & Security Enthusiast
5 个月Really amazing
Data| AI/ML | Causality
5 个月Thank you very much for this. Really insightful and awesome inspirational. I have a question concerning the second correction of the stacked bar chart. I thought it'd have been better if "unfavorable" came after neutrals ("don't know" and "neither". It seems like a more natural hierarchy, , in my humble opinion? "Favourable" being after the neutral options puts it out of place. Just curious to why it was done that way