Visualising accidents

Graph of accidents in 2007

Visualising accidents sounds like a slightly morbid photography project, but want I actually want to write about is slightly less morbid:  I finally created a visualisation to illustrator some of the data concerning causes of death, published by the Guardian Data Store. I’d still like to create a program to organise data about countries so I can do more complex comparisons, but it’s proving complex to organise all the different type of data. Maybe this will inspire me to finished it. The data I used for the visualisation is a selection from a huge data set about the causes of death for men and women in the UK in 2008 and 2007.

The selection of data I plotted is slightly arbitrary; it was determined by what I found interesting, but more importantly what icons I could find to represent each cause (for example, I’d really like to include ‘accidental drowning in bath tub’, but I can’t find a good bath icon). Most of the icons I used are from road signs, where there are free, open source SVG images I can use (e.g. at I might try to make some icons myself, although making this graph has given me a renewed appreciation of the difficulty in creating a good icon: they must be as simple as possible while containing the essence of what it represents).

The causes I plotted are:

  • Lightning
  • Contact with hornets, wasps and bees
  • Bitten or struck by a mammal
  • Accidental suffocation in bed
  • Struck by thrown, projected or falling object
  • Exposure to excessive natural cold
  • Accidental alcohol poisoning
  • Accidental drowning
  • Exposure to smoke, fire and flames
  • Exposure to narcotics and hallucinogens
  • Falling (which I’d like to separate into falling from stair, ladders, buildings and cliffs if I can make some icons)
  • Traffic incidents involving air or space (!) travel
  • Traffic incidents involving water transport
  • Traffic incidents in which cyclists are killed
  • Traffic incidents in which motor cyclists are killed
  • Traffic incidents in which pedestrians are killed
  • Traffic incidents in which car occupants are killed

Hopefully, you can tell which is which from the icons. It should also be immediately clear that males are more likely to be involved in the accidents shown, especially those involving motorcycles. More females died through bee an mammal attacks this year, but that’s not true of 2007 (not shown here). Only falling and exposure to the cold are consistently more likely to kill females.

The idea of plotting different icons of a graph is fairly new to me, and could be considered unscientific as it makes it impossible to read to true values of the graph. However, you can still get an idea of the value and could argue that is shows that there is some uncertainty in the numbers (maybe the size of the icon could be used to represent variation in the data if, say data from different years was plotted). It also gives you an additional dimension, so more data can be plotted (here I separated the data for male and female deaths)

This graph is similar to one I uploaded to Flickr:, but includes a few more causes and leaves off data from 2007, which would have made the graph look untidy. I quite like being able to see multiple years worth on data on the graph, so changes can be seen. I think it would be interesting for example to see whether deaths in car accidents have been decreasing (I think they have – but I only have data from two years). I’ve been delving into the Office of National Statistics website, to find information from other years. It’s a bit hard to find the right information under the mountains of information, but I applaud them for providing it, and think I have finally found what I’m looking for. The question is whether the causes recorded are the same each year. There is far more data available there, including the age of death, which might also be interesting if I can think of a way of displaying the information. It also includes accidents involving three-wheeled and animal-drawn vehicles.

Post new comment

The content of this field is kept private and will not be shown publicly.