Research visit in New Zealand: Modelling the spread of COVID-19 on higher-order networks

At the beginning of March, I went to New Zealand for a research visit with a planned duration of four weeks. This visit was part of a collaboration involving the IEG DH Lab and Te Pūnaha Matatini, a New Zealand centre of research excellence for complex systems and data science hosted by the University of Auckland’s Department of Physics.

The main goal of the visit was to further develop our research on the structure of higher-order networks, that is, networks which account for interactions between individuals stemming from their membership to groups. Higher-order networks are usually represented by either bipartite (two-mode) networks, or hypergraphs, or simplicial complexes. Our last publications on this topic can be seen in the Journal of Complex Networks and the Physical Review E (also available on arXiv, here and here).

I arrived about one week before New Zealand closed its borders to the world, allowing only citizens and long-term residents to enter into the country. That was also the time we had to discuss things related to our original plan. However, things quickly changed for us as for everyone else.

Te Pūnaha Matatini has been playing a major role in providing scientific information to support the New Zealand Government’s response to the spread of the virus in the country. New Zealand has made the news worldwide due to successfully responding to COVID-19, and I am proud to say that a big part of this is a result of the Government listening to epidemiologists and modellers, many of whom are my colleagues at Te Pūnaha Matatini.

Usually, theoretical research takes time to see its application to the real world. However, suddenly I saw myself involved in this scientific response, with a letter allowing me to break the lockdown as an essential worker for research purposes. As one can imagine, infectious diseases spread according to the patterns of interactions between people (as do ideas, innovation, opinions). Our recent work is about this, showing in particular that interactions of people as part of groups are crucial for understanding such patterns. In a nutshell, our findings tell us that:

1- The more we find large groups in society and the more we see two persons belonging to several groups together (e.g. siblings going to the same school, partners going to the same gym), the spread of diseases is facilitated intra-community because there exists more of what we call strong ties.

2- On the other hand, the more the number of groups per person (people belonging to several groups), we find more of what we refer to as weak ties which favour the spread of diseases inter-community.

In the specific case of New Zealand, what we are doing at the moment is trying to model the structure of the contact network (through which the disease spreads) in the entire country, i.e. circa five million people. To do so, we rely on census, education, employment, and mobility data. Our goal is to have a better knowledge of how contact patterns change according to several variables, such as location, industry sectors, and population characteristics (as age and gender, for instance). With that, we aim to provide enough information to epidemiologists to calculate more detailed basic reproduction numbers (average number of new infections generated by one infected person), instead of having a single one for the whole population.

Ultimately, our research can aid policy-making for social distancing measures on several levels and ease the job of testing, sampling and traditional contact tracing. Moreover, it can help in identifying the most vulnerable communities to give them the appropriate support against the spread of the disease.

Cite this article as: Demival Vasques Filho, "Research visit in New Zealand: Modelling the spread of COVID-19 on higher-order networks," in Digital Humanities Lab, 15/05/2020,

Schreibe einen Kommentar

Deine E-Mail-Adresse wird nicht veröffentlicht. Erforderliche Felder sind mit * markiert.

Diese Website verwendet Akismet, um Spam zu reduzieren. Erfahre mehr darüber, wie deine Kommentardaten verarbeitet werden.