2013 Summer School in Statistics for Astronomers; June 3-7 2013; Pennsylvania State University, College Park
Modern astronomical research often involves organizing vast surveys of imaging, photometric and spectroscopic data producing terabyte to petabyte databases and billion-object catalogs. An alphabet soup of time-domain surveys underway in visible light, and a new generation of radio interferometric telescopes, will soon be producing enormous datasets. While the promise is great, the scientific goals cannot be achieved with the narrow suite of statistical methods and old-fashioned labor-intensive approaches in common use by the astronomical community. Modern statistical procedures implemented with computationally efficient algorithms are essential. However, due to the structure of undergraduate and graduate curricula, U.S. astronomers are generally not well trained in statistics. Most learn elementary methods through books written by and for physical scientists and covering only a narrow range of problems, providing inadequate conceptual foundations in mathematical statistics and little guidance to vast fields of applied statistics.
The five-day 2013 Summer School for young astronomers in statistical inference will present concepts and methodologies at an intermediate level, using experienced instructors and an innovative curriculum. This is the latest in the series of intensive week-long Summer Schools in Statistics for Astronomers, initiated in 2005 by the group at Pennsylvania State. If maintained at a steady state, these Summer Schools will train about 10% of the nation's young astronomers, filling a critical lacuna in the US scientific workforce.
Observational astronomers continually confront a wide range of challenging statistical and computational problems. Statistics is needed to analyze mega-datasets emerging from powerful astronomical surveys. Due to the structure of undergraduate and graduate curricula, U.S. astronomers are not well trained in statistics. Most learn elementary methods on their own through books written by and for physical scientists. However, these volumes usually treat only a narrow range of problems, providing little guidance to vast fields of statistics. To alleviate this educational gap, intensive week-long Summer Schools in Statistics for Astronomers were initiated in 2005. The Penn State Summer Schools trained nearly 575 total participants since its inauguration in 2005. The Summer Schools seek to give a broad exposure to fundamental concepts and a wide range of resulting methods across many fields of statistics. Oriented towards graduate students and young researchers, participants in the 2013 Summer School in Statistics for Astronomers received an intense immersion in statistical methodology taught by highly skilled and experienced instructors. The Ninth Summer School has had a three-pronged curriculum: instruction in the underlying principles of modern statistics; exposure to advanced methodologies useful in astronomy; and hands-on training in the R statistical software package using real astronomical datasets. Filling a critical lacuna in the US scientific workforce, this effort should have a substantial impact on the training of many young STEM researchers. Thus addressing one of the principal goals of NSF to foster integration of research and education. The astronomical community's interest in the program is strong. The 2013 summer school was held at Penn State during June 3-7. Twenty out of the forty-eight participants were females. In addition to US participants, the researchers came from Europe, South America and Australia. Thus the summer school has had strong impact on young researchers with gender and international diversity.
