Education and training generate a lot of data still largely underexploited. They could, however, help students to better orient themselves while providing a tool for teachers to evaluate the devices more effectively and especially to adapt them to students in real time.
To answer a MCQ, to see a pedagogical video, to make a request on a search engine … Learning often passes through a surfing on the Web. However, the Internet-learner leaves traces of all his actions. So why not exploit these big data to improve education?
It is now what academics, entrepreneurs and institutions are working on. In the same way that companies such as Google, Amazon or Netflix observe Internet users in order to better understand them as consumers, it is now possible to better know the Net surfers as learners, both individually and as a collective.
Do not miss the students’ orientation
Selected for you
Even if the heatwave makes its housing unlivable, a tenant can not …
The predictive analysis of large masses of data has different applications. The start-up Cialfo, created in 2012 in Singapore, has for example developed a service called Companion which helps students to find a training that corresponds to their expectations.
Similarly, in France, the American NGO Bayes Impact, founded by a young Frenchman, 24 years old, Paul Duan, collaborates with Pôle Emploi to support the unemployed and determine according to their profile what is the most suitable path for a return to the ” employment. It is thanks to the exploitation of the large masses of data that the site Bob-emploi, launched in November 2016, was born .
Data analysis can also help recruit new students into higher education institutions. “Recent studies show that predictive analysis will determine which students will enroll in a training, who will graduate and who will become active members of the training alumni network once they enter the world of work” , Explained recently the British consultant Marguerite Dennis on a blog dedicated to education .
Selected for you
Contaminated eggs: lots delivered in France
Testing teaching methods
On a large scale, data feedback contributes to evaluating the effectiveness of an education system. “So far in education, the only measure we have is evaluation at the end of the lesson, but we can change as educational processes become digital, with MOOCs for example. How to teach that works or does not work according to the student, his age “, explained Kenneth Cukier, author of the book” Learning with big data “, in an interview conducted in April 2014 . Massive data analysis helps to design and implement more effective training schemes,
But the exploitation of data can have even more advanced applications in the educational field. Tested in several schools and universities around the world, adaptive learning is “one of the most interesting applications” of Big Data in this field, says François Taddeï, director of the Center for Interdisciplinary Research (CRI) in Paris. It is a matter of adapting the exercises in real time according to the way of learning from the person who realizes them. Indeed, the behavior of a student before a digital learning device reveals what he needs to deepen, his way of learning and assimilating.
The benefits of adaptive learning and learning analytics
Powerful algorithms make it possible to exploit in real time the information returned in order to guide the rest of the educational session towards a more tailored teaching. “It becomes possible to adapt the interval between two checks according to the successes and failures of each student,” describes François Taddeï. In exercises on tablets ( as in this report of Arte Future for example ), “we can make work together of students who have understood well with others who have not yet understood the exercise by monitoring the results in time Real by the teacher, “continues François Taddeï.
In the United States, tests have been conducted for several years and the results tend to demonstrate the effectiveness of adaptive learning. Teachers at the Arizona State University observed a 10% increase in the success rate and a drop in the dropout rate of 50%. Start-ups of the EdTech are developing, like the New Yorker Knewton, which is part of an industrialization of the personalization of education.
In France, the MOOCs Unow platform promises to adapt to learners. Similarly, Domoscio markets a learning aid software designed with advances in cognitive science and adaptive algorithms.
The data make it possible to have a quick return on several teaching methods and to choose via the iteration, that is to say the repetition of these tests, which is the best teaching method according to the context. The lean start-up is then applied to education, as explained by the developer and entrepreneur Khurram Virani, in a TEDx intervention in Vancouver in November 2015 .
Vocational training is no exception
Finally, this learning progress will also apply in companies, to develop vocational training. “It is likely that this will develop at least as quickly in companies, and in particular it will be possible to encourage the sharing of information between colleagues, for which we can imagine the matching of skills via an application, A web application inside the company will be able to detect and connect two people with complementary skills to help each other on a specific project, “François Taddeï congratulates himself.
In short, the data could well change the way we learn, whether in the classroom or in the workplace. Advances in this area are still in their infancy, and as is always the case with the use of personal data, their deployment is accompanied by a major ethical issue: “We must ensure that The learner, is the first beneficiary of the exploitation of these data, and not the GAFA, “concludes François Taddeï. A consideration that he will have the opportunity to highlight in his report on pedagogical innovation, commissioned by the Minister of National Education Najat Vallaud-Belkacem and expected in March 2017 .