Aug 23
I started to play with the openurl.ac.uk dataset http://openurl.ac.uk/doc/data/data.html - inspired by the Discovery competition run by "UK Discovery and the Developer Community Supporting Innovation (DevCSI)". Unfortunately, partly due to this being summer, I didn't get much time to work on it, and only the most bare skeleton is up and working:
This tried to provide some general usage stats based on the data. Next steps would be to include time based data, e.g. when is a particular journal/article/source popular? Secondly, a compare two or more journals/articles/etc. Thirdly graphs and charts for everything! fourthly, make it all pretty and dashboard-like. There's also potential to bring in data from other sources, and link out, especially for journal titles.
More cynically, it would be interesting to try and reverse engineer the institution resolver id to University name.
I used netbeans as a IDE, git and github for code tracking/sharing. This is the first time using all three of these, so there was a bit of a learning curve as well as just diving in to write code.
I also made a bit of a start with another idea, to provide a service for searching publishers, especially those smaller publishers who were less well known. It would then show books published by that publishers. My timeline for working on it went a little like this: see email in early July about some sort of competition. A week later actually get around to reading it properly and realise competition ends when July does, put socks on, start to work on it trying to use Cambridge and BL/Bnb datasets. SPARQL endpoint at Cambridge timed out when doing my (simple) search for publishers. BNB had no sparql endpoint according to ckan (later the record was updated to show an endpoint), and the dataset was only a small portion of bnb. Realise only have a week left to end of July and no time to make any more progress so give up (but do find a endpoint for bnb, required very different sparql query to get required data, but didn't timeout which was nice). A week later the competition re-opens, but after a little more playing decide to move on to something else. you can see the code, as it is,here
https://github.com/chriskeene/discovery1
