Interactive time series plots

It’s been taking a little longer than I’d wanted to collate estimates of cycle activity based on the National Travel Survey, partly because there are big gaps in the 70s and 80s before it became a continuous survey, and partly because it is a complex survey.

In the meantime, I’ve produced a “simple” time series plot of the total number of serious and fatal cycle casualties reported to the police from 1979 to 2019. The biggest seasonal peak saw 36 males and 3 females killed or seriously injured on July 24th 1984 whereas the smallest seasonal peak seems to have been June 20th 2002 with 13 males killed or seriously injured. The plots give the impression that the numbers of injuries have been increasing since 2002.

You can zoom in and out these plots. But what is striking is that the peak for male fatal and seriously injured cyclists seems to happen in the summer. You would assume that this is a feature of exposure and not risk. The first assumption would be that there is more cycling that happens in the summer. You would not try to explain this pattern in terms of risk; indeed you would assume the risk to cyclists is greatest in the winter when the weather and light can be poor. It is also striking that in some years there was an October peak for female casualties, possibly coinciding with half term.

The next stages for this analysis will be to quantify the amount of cycling by time of year based on the National Travel Survey.