recent

Authenticity in the age of Instagram

Use ’transformative contrasting’ to guide ideas through gridlock

How social influence drives online gift giving

Ideas Made to Matter

Entrepreneurship

How 3 MIT grads are using the wisdom of the crowd to predict crop prices

By

Tarit Rao-Chakravorti and his team at Context Insights knew they were onto something when they started checking their phones during the Pakistani general election in July.

For the past few months, Rao-Chakravorti, Timothy Reed, and Yao Li, all 2018 MBA graduates of MIT Sloan, had been building and refining their real-time data analysis tool for African crop prices.

With the election looming and their curiosity building, the team decided to use their technology to crowdsource what Pakistani citizens in 125 cities thought would happen to exchange rates for the rupee against the U.S. dollar.

The team’s median prediction was 122:1. In reality, the rate moved 6 percent — from 130:1 to 122:1.

“It was eerie,” Rao-Chakravorti said. “We think obviously it’s a one-time outcome, but it is a glimpse at the promise that this concept holds, which is that we can tap into places where opportunities are being overlooked because of a lack of data. We can do it at scale, we can do it cheaply, and we can do it by relying on distributed intelligence.”

Wisdom of the crowds

Distributed intelligence is at the heart of Context Insights, or more specifically, the intelligence and knowledge of local farmers networks in East Africa.

Using mobile phones and financial incentives, Context Insights gathers crop forecasts from farmers to predict prices.

Li said originally the team thought of selling its data to financial institutions, and while it might still explore that option down the road, the MIT grads decided instead to work with larger multinational companies “that have direct sourcing in soft commodities in Africa.”

The idea originally grew out of a class project from Rao-Chakravorti and Reed. In 2017 the two took Opportunities in Developing Economies, and their professor Tavneet Suri challenged the students to “tackle a major market failure using advances in technology,” Rao-Chakravorti said.

The pair ended up pitching a plan to crowdsource forecasts for commodities, like maize. Rao-Chakravorti said maize constitutes almost half of the calories East Africans eat, and about 70 percent of the population works in agriculture.

The idea received backing from the Legatum Center for Development & Entrepreneurship, and in January 2018 Rao-Chakravorti and Reed traveled to Rwanda, Kenya, and Uganda to research their theory on crowd insight.

“Most of the time people who come into these villages are doing basic development work,” said Reed. “We [were there] to talk shop. We really focused on understanding exactly what our potential users know about the markets. One surprising thing was that they had quite a detailed understanding on agriculture. These are men and women who have been working in agriculture their entire lives.”

By the time the two men left the continent, they’d heard from farmers and traders and the East African financial infrastructure, and had a memorandum of understanding from the East Africa Exchange to provide predictive data analytics using their methodology, Reed said.

Li, who joined in the team in the spring, said in the last four months the company has been tracking predictions in Zambia, cocoa prices in Ivory Coast, tracking maize and sorghum prices in Kenya, and has a pilot with the East Africa Exchange to help the Rwandan government with its famine reserve purchases.

In May the team won a $10,000 Rabobank-MIT Food and Agribusiness Innovation Prize, and they are using the money for incentives and research. They’ve also received funding from the MIT Sandbox.

Right now Context Insights is focused on Africa, Rao-Chakravorti said, but the hope is to expand the scope of data they’re collecting.

“We see great potential for our idea and platform across all of the developing world,” Rao-Chakravorti said. “Basically anywhere where there is a disconnect between the opportunity and the amount of data. Where we see that there’s a huge amount of opportunity but lots of growth, but not commensurate amount of data to make sense of it, that’s where we’re excited to expand to.”

Related Articles