The dangers of accident mapping

The BBC's interactive road accident map shows how data can be made to look groovy and enlightening and that there's a great deal of potentially useful data out there.

But if I've understood it right, it gives our audiences a bad steer (excuse the pun) because it doesn't seem to be unadjusted for traffic volumes.

So when it implies that the world is more dangerous during the morning and afternoon rush hours, in fact, - at least as far as we can tell from the information in the graphics - it might actually be safer.

Why? Because your chance of being involved in an accident does not depend on how many accidents there are, but by how many accidents in relation to how many people are out there.

So if there are ten accidents when there are a million people on the road, it is safer to venture out than when this is one accident when there are five people on the road. So we simply don't know if the graphs that show when most accidents occur really mean that these times are more dangerous.

Similarly, the lovely, captivating, time-lapse map of accidents over the course of a day/year which shows bright lights all over the urban areas for each accident, conveys the impression that urban areas are more dangerous. This seems not to take account of the fact that urban areas are more populated and busier. But that doesn't necessarily mean they are more dangerous.

The piece as a whole gives the impression that simply counting accidents measures danger, when what it mostly measures is how much traffic there is.

We mustn't let the huge improvement in the availability of data and the great scope for doing whizzy things with it run ahead of our journalistic ability to make sense of it.