Model predicts migrant flows, Expleo and Uni Salento win ‘Global Hackaton Onu’

Predict migration flows using a model capable of processing a large amount of data. It is the global challenge launched by the International Computing Center of the United Nations (Unicc) and won by the young members of the data scientist team of Datafactor (Expleo Italia) and of the University of Salento, who won first place at the event ‘ Global Hackaton: Data For Good ‘, an international competition that aims to find creative solutions through Data Science. The Italian project nominated for the first UN Hackaton competed with 59 universities in 13 different countries in the world, for the development of a predictive model on migratory flows in the next 5-10 years. The results of the event and details of the researchers’ work were disseminated by UNICC on its website. It all started in February, when two professors from the University of Salento received the notice for participation in the event. As professors at a university that supports the United Nations Sustainable Development Goals, Antonella Longo, Professor of Data Management & Big Data Management for Decision Making, and Gianluca Elia, Professor of Digital Business, encouraged a group of students to take up the challenge. Students – Enrico Coluccia, Francesco Russo, Ricardo Caro, Giulia Caso, Gianmarco Girardo, Marco Greco and Chiara Rucco – registered as ‘Heel of the Boot’, referring to the seat of the University of Salento, and , in the words of UNICC, “perhaps they did not know each other, but they showed a common interest in data science in the context of international humanitarian crises”. Within days, they successfully worked out the winning solution to the Hackaton challenge, ‘Refugee Crisis: Predict Forced Displacement’. The machine learning model to predict realistic shifts (in accordance with the current geopolitical landscape) and unexpected shifts (not directly related to the present scenario) developed by ‘Heel of the Boot’, highlights Sudan, Sweden, Afghanistan and Ukraine as the main countries of origin of refugees by 2024. The evaluation commission, which validated and honored the work with honors, asked for an explanation of the reasons supporting the forecasts on flows from Sweden, considered an anomalous result. Team member Francesco Russo replied that “the model built is reliable and indicates other factors, in addition to political instability, capable of influencing forecasts”. Furthermore, he added, “if a model reflects only what we have already learned from the past, then it is not a model”. UNICC’s ‘Global Hackathon: Data for Good’ took place on Tuesday 16 February 2021, and attracted a total of 140 students from 54 universities in 13 countries around the world, to tackle three of the main challenges related to the United Nations : the ‘Covid-19 Open Challenge’, ‘Refugee Crisis: Predict Forced Displacement’ and the ‘Un75 Visualization Challenge’.