An updated list of publications and working papers can be found on my scholar page.

As a result of a crippling fear of commitment, I worked on many unrelated things. My two main focus however have been non parametric statistics (topological data analysis in particular) and bayesian spatio-temporal modeling (with and without applications to public health).

Bayesian spatio-temporal models for COVID-19

During my post-doc at Imperial, I worked on the development of spatio-temporal models for different metrics of COVID-19 burden (test positivity, prevalence and mortality). In particular, we tried to assess how socio-economic deprivation and air pollution were related to the spread and severity of COVID-19, and quantify the impact of COVID waves on different ethnic minorities in England.

• COVID-19 Infections and health equalities

Padellini, T., Jersakova, R., Diggle, P. J., Holmes, C., King, R. E., Lehmann, B. C., … & Blangiardo, M. (2022). Time varying association between deprivation, ethnicity and SARS-CoV-2 infections in England: A population-based ecological study. The Lancet Regional Health-Europe. link

• COVID-19 Mortality and air-pollution

Konstantinoudis, G., Padellini, T., Bennett, J., Davies, B., Ezzati, M., & Blangiardo, M. (2021). Long-term exposure to air-pollution and COVID-19 mortality in England: a hierarchical spatial analysis. Environment international. link

• Prevalence Estimation

Nicholson, G., Lehmann, B., Padellini, T., Pouwels, K. B., Jersakova, R., Lomax, J., … & Holmes, C. (2022). Improving local prevalence estimates of SARS-CoV-2 infections using a causal debiasing framework. Nature Microbiology. link

• Combining spatio-temporal models

Nicholson, G., Blangiardo, M., Briers, M., Diggle, P. J., Fjelde, T. E., Ge, H., … & Richardson, S. (2022). Interoperability of statistical models in pandemic preparedness: principles and reality. Statistical science. link

Topological Data Analysis

The main topic of my PhD was Topological data analysis (TDA), a set of non parametric methods for describing objects through their topological features, taken to be interpretable summaries of complex dependencies. We tried to tackle two different problems: learning the topology (as the topological structure of data may be critical in understanding the underlying generating mechanism) and learning with topology (investigating whether topological features can be used alternatively or jointly with more traditional summaries of the data in order to improve performance of statistical methods).

• Learning the topology

Padellini, T., & Brutti, P. (2022). Persistence Flamelets: Topological Invariants for Scale Spaces. Journal of Computational and Graphical Statistics. link

• Learning with topology

Padellini, T., & Brutti, P. (2021). Supervised learning with indefinite topological Kernels. Statistics. link

Other projects

• Quantile Regression for Discrete Data, with Haavard Rue working paper

• Sparse Canonical Correlation Analysis, with Lavinia Amorosi working paper

• Bayesian modelling of administrative sources for small area population estimation, with Emily Peterson & many others working paper

• Wasserstein Consensus for sample size determination
• Object oriented conditional inference trees