Although the raw data is available for download
Posted: Sun Dec 22, 2024 10:06 am
Have the UK roads become safer over time? We will look at UK road safety through the analysis of road traffic crashes over the last 40 years to see if we can answer this question. In previous blogs of this style (Note 1), we have presented all the analysis in Apteco FastStats visualisations. In this post we are showing some of these australia contact data as visualisations from our new Orbit platform.
Data collection
The UK government has made data available that stretches back to 1979 covering traffic crashes in the UK (Note 2). This data is collected by the police using a form called STATS 19 that is filled in following a crash. The dataset is not a complete and accurate account of all the crashes on the UK roads as they are only the ones that were dealt with by the police. Crashes have been classified as ‘Fatal’, ‘Serious’ or ‘Slight’. Almost all fatal crashes should be included in the data, and a significant number of the serious ones, but only a subset of all slight crashes are included – namely the ones that trigger police involvement (Note 3).
I have made use of a very handy R package (Note 4) that saved plenty of time in preparing the data ready for loading into FastStats. The structure of the resulting system is such that there are 1.4 million crashes that had 1.9 million casualties and involved 2.6 million vehicles. Note that in this blog post, I have included crashes up to the end of 2019.
Data collection
The UK government has made data available that stretches back to 1979 covering traffic crashes in the UK (Note 2). This data is collected by the police using a form called STATS 19 that is filled in following a crash. The dataset is not a complete and accurate account of all the crashes on the UK roads as they are only the ones that were dealt with by the police. Crashes have been classified as ‘Fatal’, ‘Serious’ or ‘Slight’. Almost all fatal crashes should be included in the data, and a significant number of the serious ones, but only a subset of all slight crashes are included – namely the ones that trigger police involvement (Note 3).
I have made use of a very handy R package (Note 4) that saved plenty of time in preparing the data ready for loading into FastStats. The structure of the resulting system is such that there are 1.4 million crashes that had 1.9 million casualties and involved 2.6 million vehicles. Note that in this blog post, I have included crashes up to the end of 2019.