Tackling school dropout scenario with predictive analytics

Girish Kurudi

Talking to a friend of mine who until recently owned about 7 percent of his company and was seriously thinking of holding onto his investments. He observed that the company was prospering and that he could hold on for a little longer. I asked him, why hold longer, especially when company’s price-to-earnings ratio, among other factors, suggested that the company’s potential and market adoption had reached its zenith. He trusted his instinct, however the data revealed a different story.

Wonder how this story is connected with education in India? Well, being a finance professional, I strongly believe in the power of numbers, to tell true stories and predict outcomes. Data is always right. If there is one other area that could enormously benefit from the power of data, it’s India’s education system.

According to the Ministry of Human Resources and Development (MHRD), nearly 36 percent of children in India did not complete their elementary education in 2013-2014. Nearly 20 percent did not finish their primary education. To put that into context, more than half of India’s children drop-out even before they reach the fifth grade. We can research more into data, but let’s use these figures to deliver the key point on how we are educating them?

What a big data approach can do?

One of the outstanding benefits of technology in education is the power of predictive analytics. This is the process in which data collected on students – typically attendance, assessment performance, examination results, disciplinary issues, teacher comments and observations, class engagement, online learning interaction – is used to understand learning patterns, pinpoint gaps, predict performance and identify learning opportunities.

By analysing this data on a daily basis, the software assesses for strengths, weaknesses and threats such as bullying, and predicts whether students are at risk of dropping out or of poor academic performance. The software also identifies the potential for teachers to intervene. Parents and teachers are no longer reliant on the end of term report to find out if a student is performing poorly, when it’s often too late to step in.

Leveraging on the theory of machine learning, the predictive analytics software employs a model focus on prediction, based on known properties learned from the data collected on students. Through the model of supervised learning, the software is presented with example outcomes of students who have dropped out or attained poor academic results, and input data sets of these students. The goal of the software is then to learn general rules that maps inputs to the outcomes constituting to a student at-risk. This trains the software efficiently in a supervised learning setting.

Such predictive analytics will sketch a pattern of the student. It will identify students at-risk and help not only the teachers focus on the under-performing students, but accordingly, notify the parents as well. It’s a classic case of intelligent software, which goes one step ahead and even measures sentiment to understand the state of a student’s morale. India needs this kind of deep, intelligent machine learning to help keep students at school. With real time data feedback, teachers and parents will be able to make informed decisions.

Lessons from Malaysia

Under the 1BestariNet initiative, schools in Malaysia are equipped with an integrated solution allowing teaching, learning, collaboration, and administrative functions to take place through an Internet-based Virtual Learning Environment (VLE), which can be accessed in school and from anywhere else in the world with an internet connection.

The inception of an offline, mobile School Management System (SMS) application (SPS Lite) supplemented the launch of the online web-based SMS to provide 100% universal access to and usage of the system.

Users may perform their activities offline, and the updated data will automatically sync back to the central Cloud when online connection is established.

Students have made significant gains in their development irrespective of their background or location, and will leave school with a wider breadth of knowledge and world-class skills, giving them a head start towards a better future.

Teachers can develop stimulating lesson plans with engaging content, share teaching best practice to improve the way learning issues are addressed, and track the progress of each student closely and manage any problem areas swiftly.

Parents also have access to the platform, giving them the convenience of tracking their children progress, and the ability to interact with teachers.

Imagine the benefits of ensuring that every child in India at school is being supported by such an infrastructure. As custodians of our children’s future, we owe it to them to ensure that not one of them gets left behind.

Coming back to my shareholder friend, I recently checked the stock price of the company, and true enough, it had lost about 12 percent of its market value. Data never lies. If to dream of ‘Digital India’, we should start with education, bring in data-driven approach to make everyone in the system accountable, which will surely redefine India’s educational system.

The author is Head of Education and Automation Platforms, Xchanging