In case this is useful (for when I next forget and google it myself) when scanning long documents, it’s often more reliable to scan one side at a time, and then run everything through again to scan the reverse. This gives you two PDFs, the odd numbered pages going forwards, and the even numbered pages in reverse. Then it needs to be merged.
The Cabinet Office has recently published summary details of each unpublished dataset that they know about, and asked people to say what they’d use it for.
The source data from a scrape was a bit messy, as some of the original categorisation was unclear.
I’ve put something together which shows you one dataset, to see if you can use that dataset, and tweet if you (or someone you know) might be interested in it. With 10,000 new datasets, that’s a lot, and knowing what to search will be narrow ideas (and you need to have your search term included in the random short descriptions).
If you’re after a lower intensity method, this shows you one at a time, and lets you tweet if it’s useful. It’s quite addictive, and hopefully useful to the @datagovuk team.
Disruptive Proactivity