Department of Climate and Space Sciences and Engineering in the College of Engineering at the University of Michigan


Research Fellow Lars Daldorff made Big Data analytical advances

Posted: August 11, 2015

Research Fellow Lars Daldorff made Big Data analytical advances

Statisticians have combined state-of-the-art analytical techniques from the academic and business worlds to tackle the Big Data challenges confronting astrophysicists and astronomers, according to a recent press release from the American Statistical Association.

The release states that Research Fellow Lars K.S. Daldorff and Siavoush Mohammadi, a consultant with Infotrek, told a session at the 2015 Joint Statistical Meetings (JSM) in Seattle that technical advances called "automatic explorative analysis of data" have the potential to greatly aid scientists seeking to understand the universe. Researchers using Big Data will also benefit from these advances, said Daldorff and Mohammadi while presenting a contributed paper titled Novel Application of Statistical Tools for Big Data Analyses of Solar Physics at JSM.

The new analytical techniques Daldorff and Mohammadi described are being used in a study of giant magnetic loops generated by our solar systemís sun. When physicists use large supercomputers to simulate the sun, their research produces massive amounts of data. But the phenomenon of interest is usually located at a specific point in time and space, essentially creating a proverbial needle-in-a haystack situation for the researcher.

The large quantity of data has forced physicists to reduce data amounts, which they do by looking at small portions of the data at the time, making the process long and slow before true insight is found.

But what if you could scan the entire haystack at once to find the proverbial needle? That is the question Daldorff and Mohammadi sought to answer when they looked to commercial analytics solutions to explore, categorize and display the large amount of solar research project data from the plasma simulations Daldorff had conducted for NASA.

The duo uses statistical methods that frequently are used in data warehouses and by analysts at companies to study human behavior (e.g., customers) or scientific data, in this case coronal loops. These are analytical methods that combine computational power and statistics to turn information into insight. These standardized methods, widely used in the business world, suddenly find use for a completely different type of data.

"Our hope is these results can help with solar magnetic loops research at NASA and at the same time, our work will show the effectiveness of explorative analysis of data in other data-intensive fields. There are numerous possibilities for this new application that could potentially help various types of researchers - in academia, business and science - obtain quicker insights and results from their research's Big Data," said the duo.

To learn more, please visit

This research also appears on

Latest Headlines