This was a 65 minute session on Data Mining using Multidimensional Scaling (MDS). It started with a brief presentation. We will use R. This is free high quality software tailored for Statistical Programming. Open R, open this worksheet and have a play. The worksheet includes an exercise. Before you came up to Oxford many of you filled in a Google Form which asked a number of (very superficial) questions about your Interests and Preferences. I have downloaded your answers into the spreadsheet "Interests.csv". with web address http://www.stats.ox.ac.uk/~nicholls/StatsTasterDay/Interests.csv Here is a sample solution which I used to create the MDS projection below, based on the data available after our session on Thursday. People closer together in this plot have more similar interests. People who like playing sport (in green) like summer (+ symbol). People who dont like playing sport (in red) are more likely to like other seasons (symbols o,x, and triangle).
nicholls@stats.ox.ac.uk