New AI Tool Uses Climate Data to Predict Commodity Price Changes

No one can definitively predict how commodity prices are going to move, but one new artificial intelligence (AI) tool is giving it a try. 

CommodiTrack, from Virginia-based Helios Artificial Intelligence Inc., uses climate data from around the world and AI to predict the price activity of 58 different commodities as of this writing, including corn, soybeans, and wheat, and makes buy and sell recommendations accordingly. 

“What we wanted to make available for farmers was that ultimate price prediction of where we see the global commodity going, both to help inform how they manage their business and give them hyper-granular insights,” said Francisco Martin-Rayo, CEO and co-founder. “They can look at other parts of the world and our competitors outside the U.S. to see how that should inform their own buying and selling patterns.”

Besides using climate data to predict price changes, CommodiTrack allows users to review a wealth of in-depth production and climate data for a commodity’s top 10 producing countries. For example, CommodiTrack can tell users which growing season within the past 10 years most resembles a country’s current year. 

Users can access the full suite of tools for exchange-traded commodities, such as corn, soybeans, and wheat, for $99 per month per commodity, or access all commodities for $699 per month. (Through the end of March, the $699 package is on sale for $399.) 

Mark Haraburda is CEO of Barchart, a Chicago-based tracker of agricultural market data. He said the supply chain disruptions caused by the COVID-19 pandemic sparked increased interest in daily commodity price information. CommodiTrack makes data accessible that previously would have been difficult to curate. 

“When you look at commodity markets … weather is of course a major factor…. But the amount of weather data involved in any analysis is potentially just gigantic,” he said. “So, using artificial intelligence, machine learning, and big data tools to analyze that type of dataset makes a lot of sense.” 

Is CommodiTrack for You? 

While potentially helpful, Helios AI acknowledges CommodiTrack may not be for every farmer. 

“It’s most helpful, I would say, for farmers who have some discretion over when and how they’re selling their crop,” Martin-Rayo said. “You’re not going to tell a Kansas wheat farmer things about their field that they don’t already know. But what’s been really helpful is, you can tell a Kansas wheat farmer things about Minnesota wheat that they might not know, certainly Russian wheat, Ukrainian wheat, French wheat that they don’t already know. And so, giving farms of a certain size the ability to understand what’s happening [with] their competitors, that really helps to inform them in their own strategies.” 

Martin-Rayo said he sees the tool being useful for “medium-plus-sized” farms. Chad Hart, an Extension economist at Iowa State University, said he could see farmers with 3,000 or more acres using it.

A Tool, Not a Silver Bullet

Martin-Rayo said to prove CommodiTrack works, Helios AI compared price data from the past 10 years with how CommodiTrack would have predicted prices would change. The tool was correct 70% of the time. 

That said, the tool considers only climate data. A variety of factors impact commodity prices. 

“When I look at what they put together here, it is very supply driven,” Hart said. “It’s not taking into account any demand-side factors at all…. So I would argue even farmers that are using this, they probably want something else that’s helping fill in that other side of the marketplace.” 

Hart also warns that farmers may want to walk through the tool with a Helios AI representative, broker, or other adviser to ensure they understand how to interpret and use the tool. For example, when CommodiTrack offers a “buy” signal, that means it is a good time for end users to buy, and for producers to wait to sell. 

Haraburda agreed CommodiTrack doesn’t cover everything that can influence price changes but said it could give farmers an edge.

“It’s figuring out how to make the product work for you,” Hart said, later adding: “And usually, to me, it comes back to reminding folks that if you’re in farming now, if you’ve survived this long, you’ve probably survived because most of what you do works. And so, that’s why I always come back to this idea that when we’re bringing on a new tool to work with, you want to figure out how you marry that with what you already do. Because for the most part, you’re doing all right. You just want to continue to build on that.”