Networks, data and crime

By Dr Bruno Requião da Cunha

When I finished my masters in condensed matter physics in 2006 I never imagined that in just a few years later I’d be on the border of Brazil and Bolivia seizing one ton of drugs (with eventual gunfights in the process, but that is something to be told after a few pints!!). In 2009, I was recruited by the Brazilian Federal Police to work as a field operative, and it turned out that the critical and scientific thinking of a physicist was indeed very useful to tackle complex investigations. By that time, the mapping of the relationship between people investigated by international drug trafficking became increasingly popular among Brazilian Feds. The first time I saw those crime maps, I was astonished by their remarkable similarity to the abstract graphs I once studied in theoretical physics.

41109_2018_92_Fig1_HTML

The network of federal crimes in Brazil from https://appliednetsci.springeropen.com/articles/10.1007/s41109-018-0092-1

In a nutshell, I eventually got a PhD in theoretical physics studying the topology of criminal networks, I then joined UL’s MACSI as a postdoctoral researcher, and dove into this relatively new field sometimes called NetCrime or even DataCrime – i.e. the use of network science, applied maths and theoretical physics to crime related phenomena. As many researchers have shown, criminal behaviour can be mathematically modeled to some extent, and law enforcement agencies can use such models to tailor effective strategies to fight organised crime and terrorism. In this sense, even though human behaviour is hard to predict as individuals, our collective behaviour has some sets of networked properties that are much better defined. In fact, offenders not only rarely commit a crime alone but also they tend to commit it along with the same associates creating huge illicit networks. Indeed, criminal activities are complex processes and likely to depend on an underlying network of actors involved in the activities who, like most people, are part of global social networks.

AS-672593556303873@1537370208158_xl

NetCrime’s logo

Noteworthy, we have been seeing in the last few years an enormous growth in the understanding of crime networks [1-7]. As examples of such growth, the number of crime-related researches published in PloS One have increased almost five times in recent years. And since 2015, I along with Ronaldo Menezes (University of Exeter), Marcos Oliveira (Leibniz Institute for the Social Sciences), Toby Davies (University College London) and Luiz Gustavo Alves (Northwestern University) have been organizing the NetCrime symposium (https://netcrime.weebly.com/) as a satellite event to the NetSci conference. This event brings together researchers from various fields including, Criminology, Sociology, Physics, Computer Science, Mathematics, Law-Enforcement and Police Science to an open forum to discuss the role of Network Science in understanding the structure and dynamics of crime in a data-driven scientific approach – in contrast with the long-stand view of criminology as unscientific. In that sense, the number of received submissions in NetCrime almost doubled from its first edition in 2015 to its last in 2018, and names like Marco Javarone, José Ramasco, Maria D’Orsogna, Marc Barthelemy and Martin Short have figured as keynote speakers.

Policia-Federal

Colleagues from the Federal Police in action.

Bridging Network Science, Applied Mathematics and Criminology has important practical impacts on crime fighting by law enforcement agencies. I have witnessed a few of these possibilities working as a Fed in Brazil. For instance, in 2014 and 2016 my team and I at the Brazilian Federal Police have used network science tools to crack an international child pornography ring on the Tor browser [8,9], resulting in one of the largest worldwide criminal investigations on the Deepweb so far  (https://www.bbc.com/news/world-latin-america-29639241). The effectiveness of applying network science to criminal investigation has been attracting the attention of Brazilian media (https://epocanegocios.globo.com/Tecnologia/noticia/2019/04/dark-web-como-atuam-os-criminosos-brasileiros-na-parte-mais-sombria-da-internet.html) and law enforcement agencies worldwide. Here in MACSI, under the guidance of Professor James Gleeson, we have been actively cooperating with UL’s law school and with An Garda Síochána (the Irish police force) to tackle real crime networks in Ireland.

Crime is ubiquitous and percolates in the structure of modern society possibly more than any other socioeconomic issue. Transnational organised crime, white-collar offenses, terrorism, child predators rings, among other horrible acts pose real danger to modern civilized societies. To tackle such a complex problem we need to go beyond the traditional qualitative criminological approach. Perhaps, the latest frontier in crime fighting is precisely to gradually incorporate the methods of natural and mathematical sciences to it. And in that sense, UL is surely on the vanguard. As would say the famous greek historian Thucydides during the Peloponnesian War (circa 410 BC), “The society that separates its scholars from its warriors will have its thinking done by cowards and its fighting by fools”.

Bibliography

  1. DA CUNHA, Bruno Requião; GONZÁLEZ-AVELLA, Juan Carlos; GONÇALVES, Sebastián. Fast Fragmentation of networks using module-based attacks. PloS one, v. 10, n. 11, p. e0142824, 2015.
  2. DA CUNHA, Bruno Requião; GONÇALVES, Sebastián. Performance of attack strategies on modular networks. Journal of Complex Networks, v. 5, n. 6, p. 913-923, 2017.
  3. DA CUNHA, Bruno Requião; GONÇALVES, Sebastián. Topology, robustness, and structural controllability of the Brazilian Federal Police criminal intelligence network. Applied network science, v. 3, n. 1, p. 36, 2018.
  4. RIBEIRO, Haroldo V. et al. The dynamical structure of political corruption networks. Journal of Complex Networks, v. 6, n. 6, p. 989-1003, 2018.
  5. ALVES, Luiz GA; RIBEIRO, Haroldo V.; RODRIGUES, Francisco A. Crime prediction through urban metrics and statistical learning. Physica A: Statistical Mechanics and its Applications, v. 505, p. 435-443, 2018.
  6. OLIVEIRA, Marcos et al. Spatio-temporal variations in the urban rhythm: the travelling waves of crime. EPJ Data Science, v. 7, n. 1, p. 29, 2018.
  7. OLIVEIRA, Marcos; BASTOS-FILHO, Carmelo; MENEZES, Ronaldo. The scaling of crime concentration in cities. PloS one, v. 12, n. 8, p. e0183110, 2017.
  8. http://www.pf.gov.br/agencia/noticias/2014/10/operacao-darknet-balanco
  9. http://www.pf.gov.br/agencia/noticias/2016/11/pf-divulga-balanco-da-operacao-darknet-ii

Dr Bruno Requião da Cunha is a postdoctoral researcher at MACSI working with Prof James Gleeson