What’s in the new data?

Last week we launched our enhanced dataset of Railway Inspectorate staff accident investigations. We added a further 17,000 records to the existing database, available here, bringing our coverage to around 21,000 individuals, 1900-1939.

In the blog post launching the new data, we outlined a few key points and gave some brief examples. In this blog post we’ll continue those ‘headlines,’ to give you a slightly better understanding of what’s in the database. After all, 21,000 cases is a large number – with 398,588 cells containing information, that’s too much to take in at a glance. Below we’ll pull out a few aspects and try to make them visible at a glance, with a bit of context where necessary.

Before we do that, it’s worth reminding you that we’ll still be looking at ways of grouping the information. When we do that, we risk forgetting that each case within whatever combination represents a person. These will be people like 58-year old James Donnell, working for the Great Northern Railway of Ireland as a gate keeper at Lurgan in Northern Ireland. On 3 April 1925 he was in charge of the Lake Street level crossing when he stepped in front of an express train and was killed.

In what follows, we’ve done some initial analysis on the whole of the c.21,000 cases in the database from the Railway Inspectorate reports, including the c.17,000 new entries.

We’ve split the island of Ireland into 3 categories, to try to help possible searchers and reflecting the political changes that occurred around Irish independence. Each of the nations is broken down by the historical county in which the accident occurred, and again down to the specific accident location.

It’s worth remembering that this dataset only contains the c.3% of staff accidents which were investigated by the Railway Inspectors. That means there are a very large number of people not to be found in our database – though watch this space, as we’ll be including more records in the future.

Ideally things like the age distributions would be contextualised, to show how many people were employed within each age group. That way we could get some impression of the relative proportions of each group being hurt in accidents. We might expect, for example, a lot of workers to fall within the 20-29 and 30-39 age categories, which might account for the peaks we see in the chart above.

Whilst the chart above makes it appear that fatalities are the largest category – which would be a worrying state of affairs – we should remember that all of the other categories (in blue) together form the injuries. It’s also important to re-state that what we see in this and all of our charts in this blog are only based on the accidents which were investigated. We don’t know how decisions were made about which accidents to investigate, so we don’t know if some types of cases were marked out as being particular significant and therefore requiring investigation.

These categories aren’t foolproof – they’re dependent upon the transcribers having made a decision about the nature of the accident, according to the options we gave them (reflected in the chart). They do, however, give an impression of where/ how accidents happened. Workshop accidents are a very small proportion – for reasons discussed here.

This was worth including to demonstrate very visually the stark difference in number of accidents investigated between women and men. There are a great many possible reasons for this which we’ll return to in the future (though do see our existing blog posts focusing on women’s accidents).

In the charts above we’ve pulled out the railway companies forming the top 10 (perhaps it should be bottom 10, given they’re the worst statistics) in terms of casualty numbers. Given over 200 companies appear in the database, there are a lot more that could feature – but there’s also a long tail of small companies with relatively few casualties. These charts, and those that follow, don’t correct for things like size of company and passenger-miles or ton-miles, so must be treated with caution – we’re not necessarily comparing like with like.

The second of the charts above, showing the casualty figures as relative percentages, gives us a slightly different picture. We can see that whilst the Great Western Railway was the 10^th company on the list (i.e. the smallest number of casualties of this group), it was actually the worst in terms of proportions of fatalities to injuries. This may well be a statistical artefact, reflecting how the data was gathered at the time, but it needs further investigation.

The charts about the companies are also split into two: above we have 1900-1922, and below 1923-39. This reflects the state-imposed reorganisation of the industry at the start of 1923, which saw around 120 companies ‘grouped’ into four big companies and a number of smaller ones.

Of the post-grouping charts, the ‘Big 4’ companies (Southern Railway, Great Western Railway, London & North Eastern Railway and London, Midland & Scottish Railway) unsurprisingly feature at the ‘top’ of the list. They had by far the biggest numbers of staff, route mileages and so on.

The final chart was worth including as a reminder that whilst the vast majority of accidents featured in the database involved railway employees, by no means all of them did. A number were contractors or sub-contractors working on the railway. Others still were not connected with the railway by employment, but people who had some cause to be around the railway (for example, coal or timber merchants).

So – that was a brief visual look at some of the ways we can understand what the database contains. There’s plenty more in there, too – but that’s for another time (and a bit more analysis!).

In future blog posts we’ll focus more on the individuals – and we always encourage you to write blogs, too, about what and who you’re finding in the database. There’s more detail here about writing for us, and please feel free to contact us with your ideas!