clairemiller.net

With FOI think MVP

| 2 Comments

Being quite busy recently, I’ve been thinking about ways to use Freedom of Information Act requests to generate stories without taking up huge amounts of time (mine, and FOI officers)

I’ve got a couple on the go at the moment – one involves working through reams of documents to pull out information and get it into some kind of structured spreadsheet, the other involves adding a line or so of data to a spreadsheet each time a new response comes in.

Actual reality – never going to find time to get through the reams if documents to turn it in to a spreadsheet, as it involves reading the documents, pulling out the relevant data and then re-typing it into the spreadsheet. So it’s not very likely to get put into the structure it needs to be out into to find the stories.

The other one – looking at the number of cancelled operations because of bed and staff shortages at each hospital trust- is more realistic. The data comes in as a small table for each trust, which can then be copied into the master spreadsheet, about 5 minutes of work for each trust. This is a bit of a slog when lots of response come in at once (which they invariably do on day 20) but is much more do-able.

When putting in FOI requests, it’s worth thinking about what you’re trying to get back. It’s tempting to ask for all the information you can, but what are you going to do with it? Think ‘minimum viable response’ – bit like minimum viable product, think what’s the minimum I need to get this question answered and this story off the ground. If you get more, great, if the answers raise more interesting questions, you can expand on what you planned.

With the cancelled operations example, the question I was looking for an answer to was whether hospital trusts were seeing more cancellations due to staff and bed shortages over the past three years, possibly as a result of increasing pressures on services. The information needed is the the number of cancellations in each of those categories for the past three years. Some trusts may send more of a breakdown, others may give an idea of how the number of cancellations compares to the overall number of operations etc. all useful but secondary to the main question.

With a question to be answered and an idea what the responses will look like, it’s easier to build and system, and spreadsheet, to feed the data into and therefore makes analysis a lot easier than a vague-ish fishing trip.

2 Comments

  1. Pingback: How to get the best data for journalism-Passing the baton of knowledge on. | Data Queen

  2. Pingback: Data Journalism for Beginners: Where to find some data | clairemiller.net

Leave a Reply

Required fields are marked *.