By Opinno Editor de MIT Technology Review en español. Alba Casilda.

Using mass data analysis of students we can see their progress in real time. New tools can monitor their reactions, successes, and failures. This information is key to offering teaching that is tailored to the needs of each and every student.

It’s six o clock. Homework time. Instead of underlining text in a book, a student connects to their school’s online platform to watch a video on a lesson that they need to revise. Halfway through the video they stop, rewind a few minutes, and go back over one of the parts. Once they’ve watched the relevant segment, they close it without watching to the end. Before this student goes back to school, their teacher can work out if they’ve taken in the key concepts.

They will have gathered information on when the student was online studying, when they stopped or speeded up the video, or if they abandoned the video too soon. All this is made possible through the application of big data tools. Such technologies gather a huge amount of data on students’ learning processes, monitoring their reactions and forming conclusions: what areas they’re struggling with, where their skills lie, and if they’re experiencing issues with studying such as attention deficit problems.

Big data in schools

When it comes to handling this data cocktail, teachers are not alone. The tools are often accompanied by Artificial Intelligence based solutions that can offer advice and show each student the steps they need to take to improve.

These tools represent an opportunity to create teaching that is tailored to what students need at any given moment. They measure how they’re taught and how they learn, considering that handling and comparing such a volume of data using conventional methods would be almost impossible. “These methodologies make a break from the mind-set of traditional education where all students are equal. Every person takes in knowledge in a different way. But big data is no more than a tool, and in order for it to work it needs to be accepted by the education system“, explains the Director of the Escuela de Ingeniería Informática de la Universidad de Valladolid (Valladolid School of Electronic Engineering), Benjamín Sahelices.

Data, data, and more data

“Giving students a grade at the end of a school year isn’t enough, because it’s not creating enough information. We need educational models that generate data. It could be on simple things such as periodic attendance and class participation reports“, points out Sahelices. Different solutions can be used according to school age: there are apps that record students’ progress, platforms to create learning schedules to close any gaps in knowledge and develop every student’s strengths, and programmes that know what students are doing at any given time.

At present, education based on learning analytics is an emerging model. “To see examples of where it has been fully implemented we need to travel to places with flexible education systems such as Scandinavian countries, the United Kingdom, and the United States“, says Sahelices. The AltSchool in San Francisco (United States) is one of the pioneers in applying these methodologies. This school uses these methods with primary school-age children. The computers that pupils use analyse everything they write, and classrooms have cameras that record classes to interpret students’ faces and the way they talk. This all helps the school reach its goal: to create personalised plans for every student.

Personalised teaching is the main goal of big data in education

“This is achieved because schools can make use of previously unobtainable classroom intelligence. The idea is to benefit both students who are lagging behind and those who are ahead“, says Fernando Herranz, who works for the Departamento de Innovación de Santillana (Santillana Department for Innovation) and works on the A20 project. The initiative, developed by Santillana in collaboration with the adaptive learning company Knewton, created a programme for studying algebra.

The pilot project has been tested in nine countries (Spain, Mexico, Argentina, Colombia, Chile, Ecuador, Guatemala, Peru, and Venezuela). Although it is not yet ready to be marketed, it served as proof of how learning pathways can be shown in real time. “Here it’s all about harnessing the predictive capacity of technology. Based on the amount of time a student takes to answer a problem, or the number of right and wrong answers, the platform can either show them how to progress, or indicate that they need extra help from the teacher“, Herranz explains.

In the same way that big data can predict the best exercises the student needs to do to get ahead, it can also work as a radar to detect any potential risks and anticipate issues. “We can create behaviour patterns and from there establish projections for the future. If a variation is detected it means something isn’t working“, explains Lluís Pastor, Director of the eLearn Center at the Universitat de Catalunya (UOC), who goes on to say how this enables them to address one of the problems with distance learning: online course abandonment.

“In 2014 we started to use these tools and found that time was a principal factor in students not completing their studies. In this type of training users manage their own time – and they need to be taught how to manage it“, he points out. So they developed a series of formulae to detect indications of students abandoning their studies as early as possible. For example, alarms ring if at the start of the course a student doesn’t connect in the first ten days, or if they haven’t taken part in any participatory activities in the days leading up to handing in a task.”By monitoring students, tutors can intervene sooner. As a result we’ve managed to reduce the abandonment rate of graduate students by 30%”, claims Pastor.

Challenges ahead in the classroom

Although we’re beginning to see big data initiatives in education, there’s a long way to go until we see it fully introduced in classrooms. “In the first place, new learning tools being designed need to go beyond data for data’s sake. The products need to take into account methodologies and teaching materials“, says Herranz, from Santillana’s Department for Innovation.

In this vein, one of the main things we’re looking at is the design of gamification-based solutions, which give lessons through games that pose a challenge to users and connect rapidly with them. “A model to follow is that used by language learning apps such as Duolingo, which introduces new concepts when a user passes a level. But to incorporate these technologies, it’s essential that schools use these devices“, acknowledges Herranz.

Without a doubt, one of the biggest controversies has to do with privacy. To what extent is it ethical to put cameras in classrooms to analyse students’ attitudes, or computers that look at everything they write, or platforms that know their weaknesses inside out? According to Sahelices, as well as taking the law into account, it’s essential to get the whole of the educational community on board: teachers, schools, parents, and students.

Experts agree that big data is not a panacea to solve potential issues of the educational model, but rather a springboard for creating systems that are better tailored to society today. We need a system that’s ready to embrace these tools and change teaching methods. That is the only way teachers and students will be able to reap the benefits of data analysis and achieve a more efficient and personalised educational experience.