Information spread, influence and attention

We build theoretical models to investigate how content gets online attention, who adopts it, and who can influence the process.

Detecting extreme ideologies in shifting landscapes

Last November 2022, we presented to the Defence Human Sciences Symposium 2022 our novel ideology detection pipeline. In this post, we …

We spent six years scouring billions of links, and found the web is both expanding and shrinking

More than a quarter of a century since the first commercial use of the online world, its growth is now slowing down in some key …

Causal Inference: A basic taster

An introduction to the core motivation, underlying theory and practical methodology of causal inference through examples.

evently: simulation, fitting of Hawkes processes

We introduce evently, an R package designed for simulating and fitting the Hawkes processes and the HawkesN processes.

Disinformation and online problematic content

We develop methods to detect and address weaponised disinformation and problematic content (hate speech, misinformation, conspiracy theories, anti-minority rhetoric).

Opinion Market Model: Stemming Far-Right Opinion Spread using Positive Interventions

We introduce the Opinion Market Model (OMM) as a testbed to measure the impact of positive interventions on online opinion dynamics.

Doing research online (by SAGE Research Methods): A Mixed-Method to Explain How Extreme Opinions Infiltrate Online Discussions

SAGE Research Methods produced a full-length interview on our interdisciplinary approach on fighting online problematic speech.

Slipping to the Extreme: A Mixed-Method to Explain How Extreme Opinions Infiltrate Online Discussions

ICWSM 2022 paper: “Slipping to the Extreme: A Mixed-Method to Explain How Extreme Opinions Infiltrate Online Discussions”

Discovering coordinated disinformation via Hawkes processes

ECREA 2021 presentation: “Discovering the Strategies and Promotion Schedules of Coordinated Disinformation via Hawkes Intensity …

birdspotter: A toolkit for analyzing and labelling Twitter users

We introduce birdspotter, an R tool that models Twitter users’ attributes and labels them.

User Analysis on reshare cascades about COVID-19

In this tutorial, we apply two novel tools (BirdSpotter and Evently) for analyzing Twitter users on a COVID-19 retweet dataset.

The labour markets of tomorrow

Our research proposes adaptative and personalised methods to help workers transition into new occupations by accounting for their experience and personality profiles.

Acquaintances beat close friends for job connections, huge LinkedIn study shows

LinkedIn's most recent experiment shows a causal relation between weak link formation and job mobility.

Job Transitions in a Time of Automation and Labour Market Crises

We build a machine learning-based recommender system that can accurately predict the probability of transitioning between occupations.