Schedule a Demo Sign In

How Artificial Intelligence Gives a Voice to Vulnerable Migrant Populations

Two men sitting on dirt road in foreground with a big city in the background

The COVID-19 pandemic has created countless challenges in the life and governance of cities. However, these challenges are not the same for all the inhabitants of Latin America and the Caribbean. The pandemic has hit the most vulnerable in particular. Migrants are exposed to the impacts of the pandemic through their work on the front line as well as vulnerabilities related to housing conditions, poverty, and discrimination.

Big Data & the Public Sector: Migration & Social Networks

To understand the experiences of the migrant population during the pandemic, the IDB leveraged Citibeats technology to monitor social networks on the perceptions of citizens about migration in 19 Latin American cities.


This project used Social Big Data as the main data source, specifically public comments on Twitter, digital media, blogs, and discussion forums.

Data Collection:

Since February 1, 2020, real-time data collection on migration was carried out in 7 countries: Argentina, Chile, Colombia, Costa Rica, Ecuador, Panama, and Peru; and in 19 cities: Arequipa, Arica, Barranquilla, Bogotá, Buenos Aires, Ciudad de Panamá, Cusco, Cúcuta, Guayaquil, La Chorrera, Liberia, Lima, Limón, Quito, Riohacha, San José, Santiago, Trujillo, and Tumbes.

Analysis of Information:

Using an artificial intelligence algorithm based on natural language processing (NLP) and semi-automatic learning, the Citibeats platform aggregated, analyzed, and structured comments on different topics while maintaining individual privacy and anonymity. Subsequently, the data was divided into relevant categories: Housing, Urban Infrastructure, Citizen Security, Regularization, Social Organization, Education, Public Services, Employment, Health, Food Security, and Mobility.

Main Citizen Concerns: Security, Infrastructure, & Health

IDB wanted to help the region’s municipalities take appropriate measures to address the situation of migrants in real time while giving a voice to a sector of the population that is often invisible.

Once the information was analyzed in depth, a direct communication channel was established with the authorities of the countries and cities involved. The project’s findings guided their interventions, inspired their communication campaigns, and quantified and validated their hypotheses about the digital conversation.

Citizen Security:

Citizen security was one of the main concerns detected in the population. Approximately 25% of online conversations revolved around this topic. The migrant population denounced being stigmatized as a cause of citizen insecurity in cities. Thousands of testimonies collected in the investigation showed how the word “migrant” was used as a synonym for “criminal.” It is of the utmost importance to expose and communicate these findings, as discrimination creates unrest and conflict that often leads to hatred and sometimes violence. Without timely intervention, the situation can escalate and become a problem in our cities.


This analysis also identified problems involving the transit of thousands of migrants back to their countries of origin from the beginning of May to June 2020. As a result of the exodus, many migrants made settlements in public spaces in addition to saturating border crossings. This activity aggravated the infrastructure problem and, as can be seen in the graphic below, generated comments as citizens expressed deep concern about these events in the context of the health crisis and demanded government intervention.


Public Health:

At the beginning of June 2020, a large influx of comments related to the Health category was detected, as can be seen in the following graph. The comments pointed to immigrants as the main propagators of COVID-19. This conversation diminished after the borders closed.


Spanish Version