 My name is Philip Scanlon. I'm in DCU, my PhD student there, and I'm going to give you a definition. Learning analytics is about collecting traces that learners have left behind and using those traces to improve learning. So we've talked about all the different sources of data. Everybody is looking at Moodle, Blackboard and the different MOOCs. In universities, anytime a student goes into the library, checks out a book, whether they swipe in or out of the library, there are all different traces that they leave behind. But another one is the edgy wrong system, the Wi-Fi. Don't know if you can see that. This is a digital trace of students for a program, computer applications in DCU. This is one particular day, and what we're looking at here are the number of requests to the Wi-Fi system by Wi-Fi-enabled devices such as your mobile phones, etc. This is for a Monday, and you can see early in the day they actually come in and they have classes from 11 to 1. They're formal classes, and in the afternoon they have labs, and you can see a lot of people drop off during the day as the day goes on. This is actually a semester. Again, hopefully you can see that. This is the first semester. You can see the start here on the left going over. This is the first week that they're in, going Monday to Friday, and eventually you can see a pattern developing. You can see the October Bank Holiday in the middle there. You can actually see a line going up. These are Fridays. On the Friday you can see that as the semesters go on they spend more and more time on Fridays. Early in the semester they're actually not there that often. This is outside school hours, or we'd say classroom hours, 8 to 5. Going again through the semester you can see that as the semester goes on they spend more and more time outside the core hours in the college. The methodology that I'm actually using, because I want to look at where students are, what they're doing, and who they're with. So I break down the university into seven different categories, which is transit, the theatre and DCU, the residences, the cafes, the libraries, the school and the sport. And then I group them together into academic and social, because research has shown that those who engage academically and socially actually do better. So I want to identify the groups who's actually with who, when they're with them and where they're with them. And then look at the pair-wise count and the location and look at the interactions to see eventually we'll be looking at who appears to influence each other. So the opportunities and challenges. The opportunity is we're going to try and identify students not engaging and identify supports. Can we see a pattern where students actually drop out? Can we identify those students before they actually drop out? Is there a pattern that we can look back over the years? All the data I'm looking at now is historical. We're not actually looking at live data. Identify peer groups. Is there a particular group size that actually do better academically? Maybe it's five or six individuals. So if there's six individuals when they come together, former group perform very well. Should we have libraries where there's actually seating for six people? Should the canteens, all the tables in the canteens in the restaurants have seating for at least six so that those groups can form communities? And then using that information, feeding it back in and using demographics of the students. Is there a correlation between student profiles, their academic achievements and their engagement within the college? That should be me.