Ateneans emerge as finalists in ADB nowcasting hackathon

August 13, 2020
Alfonso Miguel N. Sevidal

A team from the Ateneo de Manila University emerged as finalists for the Nowcasting the Economic Impact of COVID-19 Challenge hosted by the Asian Development Bank (ADB).
The Nowcasting Challenge is an ADB-sponsored hackathon that solicits proposals on how to use big data to track COVID-19’s impact on economic activity in Asia and the Pacific without needing to wait for the release of official statistics. The major economic themes covered by the hackathon were demand, external trade and income, prices, and production.
The contingent from Ateneo, nicknamed Team ACLS, is composed of Armin Paul Allado (BS ME ’12) from the Mathematics Department, Alfonso Miguel Sevidal (BS ME ’17) from the Finance and Accounting Department, and management engineering seniors Joella Marie Consunji and Nicolas Lozano.
For their proposal, Team ACLS developed a supervised machine learning algorithm that “nowcasts” imports using port calls, container throughput, cargo throughput, and customs data. The team noticed strong correlations between their chosen real-time leading indicators and official import statistics. They also extended their model to use imports as a predictor of gross domestic product and GDP’s consumption, private investment, and government spending components.
Allado, who led the development of the machine learning algorithm, shares that “wide variety of audience by proposing two regression-based approaches: a direct and an indirect approach. The direct approach aims to nowcast merchandise imports from real-time ports data. The indirect method employs a ‘nowcasting ladder’ to go from real-time ports to customs data to merchandise imports. What makes it more interesting is how we can extend the nowcasting from merchandise imports to GDP components. Surprisingly, the results point to the predictive power of the proposed model!”

Team ACLS was among the top 35 teams in the challenge, competing with start-ups, think tanks, and academic teams globally. They were given the opportunity to undergo mentorship with a lead economist from the Asian Development Bank and to present their findings to a panel of industry practitioners. Team ACLS’s mentor was Yesim Elhan-Kayalar, PhD from the ADB.
Consunji shares that “even if it was our first time working with one another, I believe we all enjoyed and learned from the experience. Each of us had our own interests and specializations that mixed well together in our output for ADB that nowcasted trade and external income amidst the pandemic.”
For Lozano, the challenge was an opportunity “to use both my skills and passions working on such an unprecedented challenge.” “I was very fortunate to have, at my age, worked with such a stellar team on this prestigious project,” he adds.
Meanwhile, Allado says that the three months spent working on the nowcasting challenge were “quite amazing.” He shares that “from brainstorming to challenging each other's ideas to refining our proposed solutions, we have all went out of our comfort zones to produce something that we can all be proud of. Truly, the power of machine learning, especially how it can be applied to our local industry, is still in its nascent stage and presents a lot of opportunities.”
Team ACLS’s final presentation deck to the Asian Development Bank can be viewed here.