Smart schools: How study of pupils’ seating, classroom lighting will boost public education

NAIROBI: Big data is one of the technology buzzwords being thrown around these days, alongside ‘mobile’, ‘cloud computing’ and ‘the Internet of things’.

The private sector, particularly in developed economies, has been at the forefront of adopting one or more versions of these concepts to grow sales, reach consumers or cut production costs.

For the public sector, however, adoption has been sluggish, especially in developing countries, including Kenya, where as late as last year a Government department was advertising tenders for typewriters, and another for the airlifting of presidential speeches.

But things are set to change as scientists at the Nairobi-based IBM Research lab introduce cognitive computing in Kenya’s public education system.

Last year, the research centre unveiled Watson for the first time in Africa. Watson is the name given to the firm’s super computer that uses specially coded software to execute functions much in the same way a human brain does.

The computer develops its knowledge from reading vast amounts of big data that are fed into it, and using that as context, answers questions and draws conclusions to problems that humans cannot find insight into.

Using this analytical power, and image and biometric data, the team, led by education scientist, Kommy Weldemariam, is looking to change how the cogs of Kenya’s public education sector turn.

“The first thing we did was to try to understand what challenges face public schools that make them not perform as well as they may want, or hinder them from achieving set goals in the educational sector,” said Dr Weldemariam.

“Many public schools are yet to appreciate the value of data because in most cases, the kind of data they have is very unclean, takes a long time to collect and by the time it gets to the intended destination, like the ministry of Education or development partners, it is outdated and a decision has already been made.”

UNIQUE PROBLEM

Data from the ministry of Education indicates that net enrolment levels in primary schools have increased from 77 per cent in 2002 to 96 per cent in 2011, with much of the bump attributed to free primary education, introduced in 2003.

This surge in demand for education means that more than 154,392 additional teachers are required in primary and post-primary institutions by next year. Further, the student completion rate in primary school has dropped from 83 per cent in 2009 to 74 per cent in 2011 is getting worse.

These challenges make real time data and analytics to shape future education policies crucial and more of necessity than just a value add, Weldemariam said.

“We use mobile technology, which has become ubiquitous these days, to gather data, and we then use big data analytics to mine and analyse the data and provide insight to the people who need it,” said IBM Research’s Jonathan Batty.

“We are focusing on the education sector, and we are carrying out a pilot project in Mombasa County with more than a hundred schools where we are using tablets and mobile phones to collect accurate data on primary schools in this area to see what resources are needed and where.”

This is a rising trend in the field of research, where mobile devices are transforming the top-down approach to research and instead engaging the community in data collection and aggregation using their mobile devices.

“With the help of these devices and special sensors that have been embedded in them, we are able to collect data that makes decision making more effective and transparent,” said Weldemariam.

“Take the example of attendance. The teacher typically takes a roll call and at the end of the school year, we talk about dropouts, re-enrolments and overall performance.

“The problem with this form of data collection is that by the time you are aggregating and analysing the data, the student has already failed and any measures you put in place do little to help.”

Under the new system, teachers will take a picture of their class with their smartphones at various intervals of a lesson. The data will then be fed into a system where it is analysed on a real-time basis, with features like student attendance, sitting position in class, lighting and other variables analysed.

Other data points like performance will capture how students approach the learning process, how they interact with content and even single out the areas they find difficult to comprehend.

“A large part of the education process in Africa is composed of bombarding students with content without evaluating if they need it or are absorbing it.

“We have built a sophisticated algorithm, and using sensors put in tablets, we can track eye movements of what the student is learning,” said Weldemariam.

The sensors can detect waning attention and send a vibration to gently prod a student back to the learning content.

MORE ELABORATION

The system also picks up areas a student spends more time reading to draw an analysis of which content needs more elaboration from the teacher.

“This data has implications across several levels,” said Batty.

“At the classroom level, a teacher can track the performance of individual students in real time. Headteachers can track the performance of several classes in the school, and data from several schools can be aggregated for deeper analysis by county, ministry or development partners.”

However, since the researchers are dealing with personal data, concerns over privacy and how intrusive the data collection process is to the pupils have been raised.

“It is important, of course, to get the buy in of parents and ensure we do not violate the privacy of the children,” said Weldemariam.

“As for the pilot project, we are starting out with non-personal data and looking at things like attendance, what kind of resources are available and the performance of the class,” added Batty.

“This data does not require advanced levels of privacy protection, but as we scale up and include more fields, then there will be need to introduce some privacy protection.”

[email protected]