Don’t look through a telescope data at stars. Today’s astronomer is more likely online. He or she can digitally schedule observations, run them remotely using a telescope in the desert and download the results for analysis. Many astronomers consider exploring the data computationally to be their first step in science. Although data-driven science may sound like a trendy term, it is a fundamental shift in fields such as astronomy.
The Australian Academy of Science’s 2015 report found that around 25% of the research efforts of professional astronomers in Australia were now computational. Many high schools and universities still view the required skills as second-class citizens, despite their technology and engineering courses.
Computing refers to both the modelling of the world using simulations and the exploration observational data. It is central to not only astronomy, but also to a variety of sciences including bioinformatics and computational linguistics. We must create new teaching methods that recognize data-driven, computational approaches as primary tools in contemporary research to prepare the next generation.
Big Data Is The Future Of Science
The 17th-century empiricists believed that understanding the world would be possible if we used all our senses to gather as much information as possible. While empirical science is a well-established tradition, there are key differences between the traditional approach to science and today’s data-driven science.
Computers can now store a lot of data, which has had perhaps the greatest impact. This has allow a shift in philosophy, data can now be gather to serve multiple projects instead of just one. The way we explore data and mine it allows us to plan for serendipity.
Consider the search for new types astronomical phenomena. Unexpected results can be achieve with large data sets. Recent examples include the discovery by Duncan Lorimer of radio bursts and Cleo Loi, a former undergraduate student of my, of plasma tubes in Earth’s atmosphere. Both depended upon mining archival data sets that were design for a different purpose.
Scientists now collaborate to create experiments that can be use for multiple projects and test different hypotheses. One example is the 135-page book that outlines the science behind the Square Kilometre Array Telescope. It will built in South Africa, Australia.
It Is Time For Our Education System To Change
One of the most iconic images of science is Albert Einstein writing the equations for relativity or Marie Curie discovering radioactivity in her laboratory. High school is where science theory and experiment taught. This helps us understand how science works. These twin pillars are often picture together with experimental scientists testing theories and theorists creating new explanations for empirical results.
However, computation is not often mention and many important skills are still un develop. Scientists need to be able to use statistical skills to design and analysis data in order for them not only do they have to be objective but also select reliable samples. This part of maths is often neglect in university degrees. Scientists need to be more knowledgeable than high school statistics in order to ensure that data-driven explorations and experiments are accurate.
Scientists need to be able to use computational thinking to solve the problems of this age. Although coding is a great start, it’s not enough. They will need to think creatively about algorithms and how to manage data using advanced techniques like machine learning.
Algorithms For Large Data
Simple algorithms for large data sets are impossible to apply, even if you have 10,000 core supercomputers. Software can be speeded up by switching to more advanced computer science techniques, such as the KD-tree algorithm for matching Astrological objects.
There are some steps being taken in the right direction. Many universities offer degrees and courses in data science. These include computer science and statistics, as well as business and science. My online course Data-Driven Astronomy was launched recently. It teaches skills such as data management and machine learning within the context of astronomy.
The new Australian Curriculum in Digital Technologies is now part of schools’ curriculum. It makes coding and computational thinking a compulsory subject starting in Year 2. These skills will be vital, but the next step in science education is to incorporate modern approaches directly into science classrooms.
Since more than 50 years ago, computation has been an integral part of science. The data explosion is making computational thinking even more important. We can make sure that our students are ready to make the next great discovery by teaching computational thinking as part science.