RIVM Centre for Infectious Disease Control (CIb) performs diagnostics, surveillance and scientific research in the field of infectious diseases. This knowledge is important to gain insight in the prevalence and spread of these diseases, to set up control measures and to get information concerning the response on the measures taken.

At the CIb, we collect for each infected person, information on basic characteristics, his location, time of infection and type of pathogen. All these dimensions (time, pathogen, place, person) have characteristics of their own, which can influence the spread of diseases. For example, the vaccination coverage in a certain municipality, a characteristic of ‘place’, will determine how likely it is that a disease will spread in that area. Consequently, at the CIb we have to deal with multi-dimensional data to gain insight into these processes.

Analysis and visualization becomes quite complicated when analyzing this multi-dimensional data, specifically when multiple types or pathogens are involved. Let’s illustrate with an example. We have an ongoing study concerning the occurrence of Human Papilloma Virus (HPV) types in girls, in order to evaluate the effectivity of the HPV vaccination, which was installed in 2009. In this study, we have taken multiple samples of the same girl over time. In each sample, we have measured the occurrence of HPV types (there are about 30 different types, the vaccine protects against two of those). A girl can be infected with 0, 1 or multiple types of the virus at the same time and this can change over time as well (infections can be cleared). In addition, we asked the girls some questions with a questionnaire about risk factors for HPV (sexual behaviour, smoking etc.) and basic characteristics. We would like to gain more insight into the patterns of clustering (co-infection of multiple types in the same person) of the HPV types in this population. For example, which characteristics do girls share with the same clustering pattern, and which HPV types do occur in these girls? How do patterns evolve over time? Is there absence of clustering in a subpopulation?

We think visualization can help us to gain more insight into these processes. The task for the student is to design, implement, and evaluate an interactive visualization prototype system, applicable to various settings (besides HPV, also for pneumococcal disease, microbiome studies, etc.) that enables epidemiologists to obtain insights. The main challenges are to deal with the combination of many different types of data (multivariate, temporal, and hierarchical) and to enable domain experts to explore these data efficiently.