So I’ve just been to the Open Data Hack Day in Oxford which was good fun. Met some cool people, wrote a lot of code and drank some brandy.
My team was playing around with using dbpedia‘s data mixed with geo-location to find you an interesting fact about where you currently are. We had a lot of fun with it — the final results are here:
- Blue Plaque (and sourcecode)
It does some neat things. It uses javascript to ask your browser where you are, or failing that to use the wikipedia name of a city, the lat/long or use the postcode. http://data.ordnancesurvey.co.uk/id/postcodeunit/SO171BJ will give you the lat & long thanks to @gothwin.
It then attmpets to find nearby places on wikipedia which are the hometown of something. It does this by searching for things within + or – 0.2 of a latitude and longitude (I know that’s not going to be a perfect square, but meh). If it finds nothing it doubles the search range and tries again until it does.
It then gets all the things that have the city as a hometown, picks one and renders a blue plaque.
For added sillyness, if there’s a image available, it has a little proxy which downloads the image, shrinks it to no more than 300×300 to be phone-friendly, and makes it white-on-blue to match the plaque.
I stole the style of buttons at the bottom of the page from m.ox.ac.uk which is an excellent example of how to make a website to work on a phone, rather than bothing making a specific phone app.
We won ‘most creative use of data’. Some of the other groups did more worthy things like visualise arts-funding data and make useful bus timetables so forth. One group had a great idea but didn’t get very far which was linked-data top-trumps. Each site in the linked data cloud has quite a few stats so you could probably do something cool with that. Most triples, most links to other datasets, most open license… Actually I wonder if there’s a tool out there which you can feed a csv and it’ll produce you nice pdfs of top-trump cards to print-out.
Good stuff. You should ask for the actual Open Plaques data (openplaques.org), too!