Practice
The exploitation of wild species: exploiting the internet for data
We are applying several natural language processing techniques in order to expand the discovery of data on wildlife exploitation from internet sources, and assemble these into a global dataset. Our aim is to identify taxonomic and geographic ‘hotspots’ of exploitation, which could be used to inform directed policy interventions.
People: Bronwen Hunter, Julie Weeds and Jorn Scharlemann
Identities, Sexism, and Discrimination on Arabic Twitter
This research project focuses on social platforms mainly Twitter in order to examine discourses about identities, social justice, gender equality, women rights, masculinity, sexual harassment, domestic violence, as well as racism. A corpus of Twitter data created using Method52 will be divided into sub-corpora of gender, language, location, and dialect. The main language of the tweets is Arabic, however English language will be included to provide a wider scope of comparison and evaluation. This research is both quantitative and qualitative in nature. On the one hand, quantitative analysis will focus on building, classifying, and annotating the dataset. On the other hand, qualitative analysis will focus on language used by social media users. Linguistic analysis will move from micro to macro levels in terms of starting from how people make use of specific pronouns, nouns, verbs, and adjectives and going further up to see how people construct their arguments and points of views. In other words, strategies that people use in order to prove their argument such as common sense, shared sentiment, religious and social beliefs and constrains, etc.
People: Mary Mansour Fouad, Julie Weeds and Justyna Robinson
Tracking Online Hate
In this research, we make use of our Method52 platform to collect and analyse data from a variety of social media sources. We typically build a series of text classification pipelines that collect datasets, geolocate the data, remove irrelevant non-hateful data, categories data according to the target and nature of the hate, and then present the results in a interactive dashboard.
People: Jeremy Reffin, Andy Robertson, Simon Wibberley and David Weir
Tracking Disinformation
Similar to our research on tracking online hate, in this research, we make use of our Method52 platform to track the spread of disinformation on a variety of social media platforms. One distinctive aspect of this work involves the use of methods that expose engagement in coordinated disinformation campaigns.
People: Jeremy Reffin, Andy Robertson, Simon Wibberley and David Weir