Syngenta Crop Challenge in Analytics contest kicks off
 Nico Martin
October 5, 2017
  • /ACES
  • /ACES News
  • /Crop Sciences

URBANA, Ill. – Food security relies on breeding crops that yield well no matter the conditions. Every year, seed companies test new hybrids in locations across the United States, in hopes of identifying the best out of thousands of hybrid/location combinations. Syngenta, together with INFORMS Analytics Society, is now asking for help in the form of a contest.

“In the last two years, I have been a witness of the impact of the Syngenta Crop Challenge,” says Nicolas Martin, assistant professor of bioinformatics in the Department of Crop Sciences at the University of Illinois and chair of the Crop Challenge prize committee.

“With hundreds of participants and submissions from all over the world from teams with backgrounds in engineering, computer sciences, and economics, they provide unparalleled insights to solve complex problems in agriculture." Martin encourages everyone with interest in big data to give the challenge a try.

This year’s challenge is to predict the performance of new hybrids in untested environments, using data from past years. Contestants download and analyze real data – some 4.8 million data points from eight years of Syngenta’s hybrid tests in over 2,000 locations – and use that to develop predictions about hybrids they tested this year.

Martin says there is no one right answer. The judges determine whether answers are reasonable from a biological standpoint, but they also take into account the methods, use of creativity, and clarity of the presentation. He calls it an academic exercise and points to the many scientific papers that have been published by past contestants.

“The contest is unique because it’s making proprietary data from the company available to the public, and allowing the winning group to retain intellectual property rights to the solution and to publish their work. Syngenta just provides the money, the challenge, and the data,” he says.

It may be an academic exercise, but finding a way to predict new hybrid performance in untested locations would be a huge boon for the agricultural industry and food security into the future.

“If corn breeders could accurately predict the performance of each hybrid in untested environments, they could make better decisions on which hybrids to move forward and provide to growers,” says Greg Doonan, head of oilseeds genetic projects at Syngenta. “With this competition, we’re tasking participants to apply their analytical and mathematical modeling skills to help solve this question. It is a great opportunity for those outside the ag industry to contribute their expertise to help deliver on the 21st-century food production challenge: growing more food to meet the needs of a growing population.”

Data and more information are available at Submissions are due Jan. 11, 2018. The first-prize winner takes home $5,000. The Informs Analytics Society sponsors the contest, and it is made available by IdeaConnection.