In the presentation you've prepared for the symposium you say that the UOC is the heaven and hell of data scientists...
Yes [smiling]. It's heaven because at the UOC everything that students do leaves a trace: when they log onto Virtual Campus, what they do in the classroom, how they access the materials, when they write a message, how long it takes to receive a reply... All this information and much more are recorded and later analysed.
In this respect, we have richer and more relevant information than on-site universities. But it's hell because we have to be very cautious when deciding which improvements to make and how.
Technologically, getting a campus that manages over 50,000 students to take on new functions is a complex task, because it can have a negative impact on students as a whole.
In other words, this action has to be approached very carefully.
Yes, before changing anything on the Campus we have to weigh the advantages and risks, and it has to be done in a very organized manner. If we haven't clearly identified the advantages and seen them as relevant - and not just for one classroom - implementing the change can be complicated.
Today there are a myriad of techniques and we have the technology to analyse all this data.The problem isn't what processes to follow but how to decide which of the improvement tools implemented are really pertinent.
How can the data help the University with the different processes?
They help us with teaching and management and, for the student, with learning. By analysing what is happening on our Campus we can see what processes are underway, how they work, and how they can be improved and streamlined. We can know what's going on in the classroom and even detect cases where students are suffering stress or anxiety.
It also tells us how students access the materials, which ones are used more, which concepts are more difficult to understand or which points of the study environment cause the most conflict.We can offer each student more personalized methodologies and content. If necessary, we could even make personalized adjustments to the delivery dates of the learning activities.
If we have all the information about how the learning process works, the sky's the limit in terms of improvement.
Next Generation Student Success Symposium
You'll be presenting many of these projects at the Next Generation Student Success Symposium. Can you sum up the main ones you'll be talking about?
We've done an initial classification according to who these projects help and for whom they generate knowledge and have come up with three: we have projects to support faculty, students and administrative staff so that the University carries out the processes more effectively.
There are projects of all kinds (gamification, forecasting, visualization, etc), and it's hard to say what the main ones are.
I'm particularly fond of the Treball@ and Orient@ projects, which are the result of an adventure we set out on in 2014 with professor David Bañeres, funded by the eLearn Center.
When we started, few people believed that what we were proposing was viable:obtaining information about the knowledge demanded by job offers (based on advertising websites), information about the knowledge developed on the courses (based on their materials and practical activities) and comparing this information to find out how far the UOC's range of training (master's degrees, bachelor's degrees, courses...) matched the demands of the labour market. And all this was done automatically, using data analysis and language processing techniques.
Finally, we created dashboards to graphically and interactively see how far we teach what the job market demands, and also to define the skills students should have and the courses they should enrol on for a specific job.
There is also a project based on the analysis of emotions...
Yes, we have feelings thermometer project that analyses the emotional climate in the classrooms. It uses feelings analysis, which is now very common, but not so much at individual student level as in general, because we believe that each classroom has a life of its own.
I'll also be talking about the Booter project, which aims to help first-year students to change their habits so they can approach e-learning more confidently by using a robot. The robot is a kind of Jiminy Cricket that supports, motivates and helps students when getting started.
One of the star projects is Tesla.
Yes, it's a European project that provides tools for completely online assessment (http://tesla-project.eu). Until now the UOC model has obviously been online, but eventually there is a point where we have to "meet" the students in person, namely the final exam.
Tesla wants to break this barrier down, creating an environment in which we can reliably identify students online, verifying their identity and that they are the authors of the online tests. To do so, we apply existing research in the field of facial recognition, voice recognition, typing rhythm and anti-plagiarism systems.
Right now we're conducting large-scale pilot tests of this project in different universities, including the UOC. By the end of the year we hope to have tested the tool with over 15,000 students.
How do you test the projects at the UOC? Do students need to be involved?
Yes, normally the project determines the courses to be tested. In the case of Tesla, given the scope of the project, we are carrying out a large-scale test with thousands of students. With other projects, such as the feelings thermometer that I mentioned, we do choose a more controlled environment. In this case, we have chosen four courses in which to test it. Students are always informed and, if they don't want to participate, we find solutions, such as moving them to another classroom where the test is not being done.
Is there a university of reference in data research and analysis whose work can be applied to improve the processes?
The Open University in the UK and the Open Universiteit of the Netherlands carry out a lot of research, but I'm not sure how much of this research ends up being applied to their campuses. In fact, if you ask me about universities that apply research and carry out improvements to their online campuses, there really isn't a lot of information and I think that the UOC could be considered a benchmark.
For example, the Treball@ project, which analyses professional employment opportunities in relation to the University's training programmes, was a pioneer and when it started no other university was doing it. Now other universities are working on it.
How many people are involved in these projects?
Some projects have only one person but others can have much bigger teams. For example, 18 universities and over 70 people are involved in Tesla.
Can the projects be marketed?
Yes, in theory. Tesla is an example. It began many years ago as an innovation project that became a European project with the objective of achieving a range of services that can be offered to society: its objective isto transfer this technology to society.
What visibility and positioning can the symposium give to the UOC?
The participants are top-level people from diverse universities and it will be a very good opportunity to publicize the tools and services developed here and to identify all the services students will need in the future. Organizing the symposium was an initiative undertaken by DxTera (http://dxtera.org), but members from other leading universities will also attend.
As a member of DxTera, the UOC forms part of a group of universities of reference that are pioneers in providing the technological tools and services students need. And not only now, but in the future. Moreover, as a European hub, it is a benchmark for all those universities and organizations that, in one way or another, wish to form part of DxTera.