In times for the holiday season at the end of November USA Today published an article calling out the worst drivers in the USA by state, based on data from CarInsuranceComparison.com.
The USA Today article provided some visualizations of the data and our challenge this week was to help tell the story more clearly.
As I used to rush to create maps whenever there was geographical data involved in the past, I decided to try my best to steer clear of maps this time.
Let’s first look at the original, though:
- there is at least some level of contrasting colours in the colour gradient chosen
- the colour legend re-iterates that lowest scores mean worst driving behaviours, while highest scores mean best driving behaviours overall.
- they thought of calling out the small states separately
- the data source is listed
- using a map helps people easily identify the areas they care about because it is a familiar way of visualising state level data
What doesn’t work?
- The heading is misleading: the data contains scores for driving behaviour, what does ‘worst states for driving’ mean? This could be related to the weather, the road quality, how interesting the area is, etc. Clarity is needed
- the colour gradient from black to blue to me doesn’t trigger any association with good vs bad. Maybe a simple black to grey would have worked better
- the additional legend for the small states on the east coast looks like another colour legend or so, which is confusing
- the heading and colour legend take up a lot of space compared to the actual data
What did I do?
I ended up using a hexmap, even though I wanted to find a ‘non-map’ way to visualise the data.
I went through an iterative process, like every week, and in the interest of time decided to just focus on the overall scores and ranking of states, rather than creating a dashboard with ALL the information.
During my analysis I looked at different charts to indicate state scores/ranks and somewhat by accident came up with the bar chart you can see above.
Originally I also planned to build a long-form dashboard, but decided against it because it didn’t suit my minimalist approach for this week.
Hex maps are something I’ve wanted to play with for a while and I think it’s suitable this week. A map showing the individual states is appropriate because the audience for this kind of information is the general public. They would easily relate to a map that shows their own state and those they are interested in. Hex maps have the benefit of showing each state with an equal size, so that no matter how big or small the geographical area of a state is, every single one is equally visible. This helps with comparisons.
So I built my hex map and added a couple of call outs. The tooltips are minimalist, as well, because I’m just looking at overall scores and there is no need to over-complicate things.
I added a title and subtitle which I believe are more specific and descriptive of the topic being visualised. My colour choice has seen me go back to one of my favourites: red to black, black to red and all the shades in between. The font is Tableau Light, because I really like skinny fonts and white spaces.
So I hope this makeover will be more useful than the original. I tried to address my critique of the USA Today visualisation but there are still plenty of different angles I could take for this story. In the interest of time, however, this is it.