08.07.06

Your Netflix Ratings

Posted in Application, Media, Movie at 10:10 pm by Robert Keogh

I really love my Netflix account. I’ve been using it for years and at this point I’ve rated almost 3000 movies. Granted just over 900 of those I never, ever want to see and most of those I wish had never been made - but hey that’s enough about me. I enjoy toying with data like this, amalgamating it and proffering it for view by friends and strangers alike.

The other day I decided to abstract my ratings data from Netflix and a cursory Google search turned up some previous work by John Ressig and Devanshu Mehta. Naturally, it was not going to be as straightforward as downloading someone else’s hard work and running a script. Netflix has changed the Login URL and Ratings page URL, as well as the HTML format of the Ratings page since these packages were written. So I figured it was time for me to try my hand at updating the requisite code.

I decided to add a few enchancements such as abstracting the URLs and regexp to a config file. Provide the functionality to capture the retrieved data in a database (PostGreSQL) if the user so desires. Cleanup the codebase and apply several of Damian Conway’s best practices.

I highly reccomend downloading Devanshu’s package as it has several python scripts that garner additional meta-data from Netflix. Devanshu has also gone to the trouble of documenting many of the basic steps required to get your environment set up, check out his blog article for the instructions. The file my code generates is in a format compatible with his. Download and have some fun.

08.03.06

Orange Hall Attacks - Timeline

Posted in Application, Conflict, Ireland at 3:28 pm by Robert Keogh

The MIT software development group Simile have created several interesting semantic web applications of which their Timeline javascript application recently caught my eye. I am very interested in alternate representations of data that yield a more holistic overview allowing the detection of patterns and relationships that are not easily visualised in flat text tables and the various bar/pie charts that hithertofore have been the gold standard. I want to be able to navigate the data in a manner akin to a mini-interweb, observe the types and nature of the relationships between related and disparate data sets but I get way ahead of myself.

Of course, one really needs sets of data to get realise this goal and boy is garnering that data tedious - as illustrated by the abandonment of my CDDB project soon after code completion.

Thankfully there are many people out there generating datasets of interest and making them available. One of those is a list of Orange Hall attacks during the course of The Troubles that Fair Deal a blogger from Slugger O’Toole has compiled. I have taken this database and ported it to Simile’s Timeline application Timeline of Orange Hall attacks.