29/07/2021

Luiz Gustavo Nonato: Data Science in the Fight Against Violence

Crédito: IEA/USP

Mathematics has multiple practical uses in everyday life, and data science has been demonstrating its applicability in the development of effective public policies. This is one of the research topics of Luiz Gustavo Nonato, coordinator of a project that aims to identify crime patterns and create tools that allow for the prediction of new cases in the future. The professor from the Institute of Mathematical and Computer Sciences at the University of São Paulo (USP) will participate as a speaker at the 33rd Brazilian Mathematics Colloquium .

Held traditionally every two years since 1957, the Brazilian Mathematics Colloquium is preparing for its first virtual edition due to the Covid-19 pandemic. The event is the most comprehensive scientific gathering of the Brazilian mathematical community, with the participation of undergraduate and graduate students, as well as leading researchers from Brazil and around the world. Registration is open, and the event will take place between August 2nd and 6th, with broadcasts on YouTube and Zoom.

Nonato will present, next Wednesday (04), at 1 pm, the dissemination lecture “Data Intelligence for Public Security”. He will demonstrate that numerous factors, such as traffic and weather conditions, population density and socioeconomic conditions, influence the number of criminal occurrences and how the analysis of this information can help in the construction of more efficient public policies. In this interview, the researcher talks about his academic trajectory, his research object and the importance of congresses such as the Colloquium to disseminate knowledge about mathematics.

IMPA : At the Colloquium, you will be giving a lecture on Data Intelligence for Public Security, a topic that is also one of your research projects. How can science help in the development of public policies to combat crime?

Luiz Gustavo Nonato : Nowadays, worldwide, public policies are guided by evidence. In this context, the field of data science plays a fundamental role in detecting patterns, identifying correlations between variables, and enabling the creation of public policies based on data and the evidence derived from that data. This is what we have been doing in the area of crime, mainly in São Carlos, in the interior of São Paulo, and also in the capital. The project seeks to understand the relationship between external variables and crime in order to propose interventions in these variables, for example, the construction of some type of urban infrastructure or the modification of the road network that has an impact on crime.

IMPA: Why did you choose mathematics?

LGN: My inspiration came from a professor at the Pontifical Catholic University of Rio (PUC-Rio). When I was in my final year of high school, I attended a lecture by Aristides Barreto on mathematics, especially applied mathematics, which greatly motivated me to pursue this path. I even took the university entrance exam for computer science, but with the professor's influence, I ended up choosing to study mathematics.

IMPA: What has your experience been like over these years as a researcher and professor, especially in the field of data science?

LGN: I started working in the field of computer graphics, and my doctoral thesis was on computer graphics and geometric processing. But over the years, especially when I did an internship at New York University, I came across the field of data science, which was still practically unknown in Brazil and was just starting there. At the time, they were creating the Center for Data Science, where I ended up working as a visiting professor. It was great to have contact with this emerging field and the curricula they were building. I became very interested in the problems this area presented and decided to migrate from visualization and computer graphics to data science. Today, what I do is basically a mix of visualization and data science.

IMPA: Your lecture at the Brazilian Mathematics Colloquium will be online, as will the other activities of the congress. What have been the challenges and advantages of holding such events virtually?

LGN: I think the main disadvantage of an online conference is the lack of human connection. We no longer have contact with people, and this personal interaction is very useful, even for establishing partnerships. But this type of event also has a great advantage, since it allows more people to access information, follow the lectures, and understand what is happening in the area the conference covers. In the case of the Brazilian Mathematics Colloquium, many people will participate online, so this ends up bringing much more visibility to the topic and opening the way for the participation of people who could not attend otherwise. And I think that, in the future, conferences will no longer have the format they previously had, entirely physical, where only people who could travel could participate. The trend is for events to be held in a hybrid format, with online activities and also the possibility of this in-person interaction.