From molecules to the Milky Way: dealing with the data deluge
Most people have a few gigabytes of files on their PC. In the next decade, astronomers expect to be processing 10 million gigabytes of data every hour from the Square Kilometre Array telescope.
And with DNA sequencing getting cheaper, scientists will be data mining possibly hundreds of thousands of personal human genome databases, each of 50 gigabytes.
CSIRO has a new research program aimed at helping science and business cope with masses of data from areas like astronomy, gene sequencing, surveillance, image analysis and climate modelling.
The research program, which began this year, is called 鈥楾erabyte Science鈥� and is named for the data sets that start at terabytes (thousands of gigabytes) in size, which are now commonplace.
鈥淐SIRO recognises that, for its science to be internationally competitive, the organisation needs to be able to analyse large volumes of complex, even intermittently available, data from a broad range of scientific fields,鈥� says program leader, Dr John Taylor, from CSIRO Mathematical and Information Sciences.
One aspect of the problem is that methods that work with small data sets don鈥檛 necessarily work with large ones.
An aim of the program is to develop completely new mathematical approaches and processes for scientists in a range of disciplines to further their research and boost Australia鈥檚 position as a world science leader.
鈥淟arge and complex data is emerging almost everywhere in science and industry and it will hold back Australian research and business if it isn鈥檛 dealt with in a timely way,鈥� Dr Taylor says.
Countries like the US also recognise the challenges, as Dr Taylor has seen first hand in his ten years鈥� working in laboratories there.
鈥淭his will need major developments in computer infrastructure and computational tools. It involves IT people, mathematicians and statisticians, image technologists, and other specialists from across CSIRO all working together in a very focussed way,鈥� he says.
After a workshop in September, specific research areas have been identified and projects are progressing in advanced manufacturing, high throughput image analysis, modelling ocean biogeochemical cycles, situation analysis and environmental modelling.
Source: CSIRO Australia
