Fifty students from universities across the country participated in an exhibition illustrating how data science can address major challenges facing the country
Despite improvements since 1994, South Africa’s mines remain one of the most hazardous working environments in the country, with seismic activity and rock-bursts resulting in numerous injuries and deaths — in 2016 alone there were 73 fatalities. Preventing this and creating a safer working environment is critical to this multi-billion rand sector.
Early warnings and monitoring rock stability in the mines could reduce these accidents significantly, which could help sustain an industry that is a cornerstone of the country’s economy.
A group of young data science students could have the answer, with the development of an early warning and monitoring system called “Rock Pulse”.
The students are part of a group participating in the Data Science for Impact and Decision Enablement (DSIDE) programme, which is funded by the department of science and technology (DST).
DSIDE is implemented by the Council for Scientific and Industrial Research (CSIR) and is aimed at building capacity in data science by recruiting students to participate in learn-by-doing problem solving to meet real-world needs.
At least 50 students from across the country were recruited in this year’s programme, which has trained 141 students since inception in 2014. Projects that fall under the DSIDE initiative are on display at the DST.
Rock Pulse works by collecting data in real-time underground, using a geophone (a device that converts ground movement into voltage, which may be recorded at a recording station) connected to a roof bolt in a mineshaft. Algorithms then interpret the data in order to identify triggers or potentially hazardous events, eventually setting off an alarm that initiates the evacuation of affected sections of the mine.
Nicolene Roux, Boitumelo Mahlobo and Clodita Mandlazi, the students who developed the Rock Pulse dashboard, confirm the technology has attained a level of confidence, which assures its validity, but that improvements are still needed.
“When designing the dashboard, we realised that the technology could not differentiate between a man-made and a non-man-made sound. This is where it needs to improve, because you don’t want a false trigger that negatively affects a mine’s production,” says Roux, who spent three years underground during her geology studies.
John Isaacs, research group leader of the Embedded Intelligent Systems unit at the Meraka Institute (a business unit of the CSIR), says South Africa’s mines are old and as resources become depleted so mines go further down, and the more dangerous it becomes.
“We need very robust technologies to be able to improve working conditions in these mines. The ultimate goal is that the system should be able to inform future planning, identify the risks and define the shortest exit paths in cases of evacuation,” says Isaacs.
He believes the system is a step in the right direction to replace the old technologies currently in use. With the country’s gold reserves predicted to be depleted by 2030, mining will become more and more dangerous, and it is systems like Rock Pulse that will create better working conditions in South African mines.
Exploring ways to increase youth employment
Other projects displayed at the DSIDE included the “Municipal Money and Youth Explorer”, which uses data to profile employment and service delivery in municipalities around the country.
Youth Explorer uses data collected in Census 2011 on challenges facing the youth, which enables researchers and policy-makers to focus on areas of concern and implement interventions. The project found that in households where no one has a matric the chances are unemployment are high — the young people in that household will likely be unemployed.
According to Thabang Mashinini, a big data and analytics student, the system is a useful tool for municipalities to improve their efficiency and provide platforms to improve employment in the country.
Detecting electricity fraud
With electricity theft a significant problem that results in revenue loss and increased costs to paying customers — as well as a range of safety issues — identifying irregular usage patterns and detecting potential electricity fraud has become paramount.
A project carried out by Advance IO, which installs smart meters for households in Ekurhuleni Municipality, is trying to identify irregular usage patterns and detect potential electricity fraud.
An electronic engineering student at the University of KwaZulu-Natal, Mixo Ngobeni, says the system is able to spot anomalous trends, which may be a result of electricity theft, meter bypass or a resetting function.
However, more still needs to be done to accurately spot the problem, improve the performance of the smart meters in the municipalities, save electricity and improve revenue.