Identification of key nodes in Temporal Networks

Network science aims to construct data representations in terms of nodes and connections among them, such that structural properties naturally emerge and are simple to characterise.
In multiple settings nodes and connections may change their properties over time, resulting in temporal network representations. In blockchains and DLTs there are multiple examples fitting this description, like cryptocurrency transaction networks - where nodes are wallets and connections are transactions happening in a given time window - or peer-to-peer networks that are the space where consensus is maintained and are periodically updated. 
The dynamic identification of key nodes is then crucial to monitor and characterise these systems, as they may be targets or sources of attacks against their integrity.
We develop methodologies that are specifically tailored towards the detection of central nodes in
- transaction networks of cryptocurrencies, to characterise intermediation and the distribution of tokens among entities
- peer-to-peer consensus networks, to identify liabilities and attack points in auto-peering mechanisms
- social networks, to spot market manipulations and develop profitable trading strategies.

The research is partially funded by the Swiss National Science Foundation through grant #200021_182659