Image: Shutterstock
Our growing ability to combine, share and make sense of large and complex datasets is opening the door to huge new opportunities for Australia and the world. Big data is powering climate and weather models, and leading to more sustainable farming with improved yields.
These two examples are just a tantalising glimpse of the possibilities. There are still plenty more big data applications to discover. At the same time, mathematicians are working hard to reduce the energy use of this power hungry beast.
Messy weather
In March, Tropical Cyclone Alfred threatened the coast of Brisbane. Thanks to data and modelling, weather forecasts provided Brisbanites with plenty of time to sandbag their homes and prepare for the huge amount of rainfall.
“Continued development in these modelling systems, through innovative approaches, can lead to more accurate predictions of weather systems such as Tropical Cyclone Alfred beyond one week [in advance],” says Savin Chand, an associate professor of applied mathematics and statistics at Federation University.
Weather data is particularly difficult because it doesn’t come neatly in one form. Modelling relies on data from different sources and types, including temperature, pressure, windspeed and rainfall. Our ability to combine heterogeneous data types and sources remains one of the most pressing challenges of big data.
From data predictions to dinner plate
Weather data is also vital to farming and agriculture, where it combines with datasets from genomic analysis, soil and sensor readings, satellite imagery, biosecurity markers, animal trackers and even market trends.
But generating all this data is just the beginning, says Professor Neena Mitter, biotechnologist and deputy vice chancellor associate of global research at Charles Sturt University.
“The real power is in how data can be integrated to make smarter decisions,” Mitter says. “How do we apply this data for solving real world problems?”
Bioinformatics is the discipline at the interface between biology and computer science. Mitter says bioinformatics is already unlocking huge possibilities for food security and sustainability.
“We once had to rely on long breeding cycles and field trials. We can now use genetic data to predict which plants or animals are more likely to thrive in certain conditions, long before they reach the paddock or the farm.”
A growing energy challenge
Thousands of data centres across the world churn through energy – perhaps 1% of the world’s total – and data is only going to become more important. But mathematics is finding a way to cut back on this energy consumption.
Professor Jacqui Ramagge, executive dean of science, technology, engineering and mathematics at the University of South Australia, and incoming ACDS president, says such a vast amount of energy is used because many computations are done by “brute force”.
Ramagge uses the analogy of a curved line graph. “Suppose you want to know where the line crosses the x-axis, where y is equal to 0”. Computers will find it by looking at every point along the x-axis and evaluating whether y is zero. A human would be far more efficient: they would use a quadratic equation.
In an effort to cut computing energy use, mathematicians are working with computer scientists to develop formulae to help computers find solutions more efficiently. Like using a quadratic equation, only far more sophisticated.
As we increasingly turn to modelling and computation to help us solve global challenges, reduced energy use from computing centres will cut down on electricity bills and also reduce the climate impacts of big data.
Written by Sara Phillips and Cristy Burne