 Good morning, I'm David Adcona, I'm a PSD student in Insight at DCU and I'm working on education analytics in computer science degrees or in programming contexts in order to help students that are having difficulties with these courses. The non-progression rates in computer science are typically double the average dropout rate in college. We design predictive models in order to predict students at risk using resource aspects logs in a similar way as you are using loop or model and we are using as well programming skills acquired in a weekly basis of what programs students write in this laboratory work they do. The features are fairly simple, programs students write and slide visited in a similar way as a model and we carried out retrospective analysis in DCU in programming one and programming two courses taught at the start of the degree in computer science where there are around 150 students and at the moment we are running real-time predictions on them and I think they are doing at this time the lab exam, the last one so hopefully my predictions work well or they do well and everybody passes, okay. And so what opportunities and challenges do we have here at the moment there is a skills gap in Ireland and a high demand from graduates in the ICT sector. Our predictive models have proven to work well with significant lab work or programs and we are offering as well personalized feedback or some kind of recommendations for those students which are being taken at the moment. In terms of the challenges as you probably know gathering data from different sources at university is a real challenge, it's a long and tedious process, you have to get the progress from different committees and students also come with low math and living cert entry points and that's a real challenge because those students don't perform well in programming and math related courses and as well programming is not that easy and only a few students find it easy at first. Those next steps we are going to take, I'm combining different data sources like Philips location data, academic history, demographics from the students in order to help those students early on in the semester as Emma from UCD was mentioning to identify the students at an early stage so we are able to help them faster. We are also looking at other courses, programming in second year and math related ones and program wise as well we want to identify who are the dropouts in computer science. We are applying different data mining and machine learning techniques as well like the recommendations we are tailoring from higher performing students in the class that are recommending programs automatically to at risk students that might keep them more motivated. So that's it, thanks very much.