The dashboard aims at monitoring extremisms in the social media network Twitter. In particular, we focus on four types of extremisms: religious, far left,far right, and separatism. In addition, we are interested in analyzing the narratives of these extremisms, as well as the counter-narratives and alternative narratives. The final goal is understanding extremist dynamics with the aim of developing social and educational interventions and policies to prevent it.
The information shown in the dashboard is collected based on manually selected hashtags for every type of extremism and narrative type. (Linguistic Inquiry and Word Count) Social messages have been processed using the linguistic resource LIWC (Linguistic Inquiry and Word Count). LIWC provides a dictionary to identify latent emotional and psychological elements in discourse in texts.
For every extremism and narrative type, the following information is shown:
Volume: number of messages per day
Polarization: positive and negative emotions of the messages.
Topics: main topics (hashtags) of the messages.
Drives and needs:
Power: impact and influence in others, associated with aggressivity in extremisms.
Reward focus: references to rewards, incentives, positive goals.
Affiliation: the concern for maintaining relations with others.
Risk: dangers, concerns, things to avoid.
Achievement: reference to success or failure,
Social: references to family, friends, male persons, and female persons.
Personal and religious concerns: references to religion, work, leisure, and home.
Geographical: geographical distribution of the messages and references to places.
Chrome Plugin
The Chrome Plugin enables users to analyze the texts they see in their browser
Making Sense of Language Signals for Monitoring Radicalization, Óscar Araque, J. Fernando Sánchez-Rada, Álvaro Carrera, Carlos Á. Iglesias, Jorge Tardío, Guillermo García-Grao, Santina Musolino and Francesco Antonelli, Applied Sciences, 12 (8143), 1-27. Available at link
Ontologies
One of the main tasks within Project Participation is the annotation of text with several types of information: emotions, narrative, ideology, etc. To that effect, and in alignment with the Linked Data principles, the project uses a combination of different ontologies dedicated to specific domains.
A distinction can be made between those that have been defined as part of the Participation project and those that have not. Likewise, a distinction can be made between those that have been developed by the group and those that are external.
Developed ontologies:
Participation ontology:
SLIWC (Semantic LIWC): It reads a given text and counts the percentage of words that reflect different emotions, thinking styles, social concerns, and even parts of speech. In the Participation project, we have produced a semantic version of the LIWC annotation schema. It consists of two parts. First, an ontology that represents the general concepts used in LIWC annotation (e.g., dimensions, categories, word-level dimensions, document-level dimensions, etc.). The second part uses these concepts to provide elements specific to the LIWC dictionaries, such as specific categories and their hierarchical relation to one another.
Morality: The morality ontology includes categories in the Moral Foundation Theory aligned with the LIWC annotation format (through the SLIWC ontology, in this case).
Senpy: The purpose of this ontology is to provide a model for the data returned by Senpy, allowing the service to use the same concepts when annotating texts using different sentiment analysis tools.
Marl: Marl is a standardized data schema (also known as an "ontology" or "vocabulary") designed to annotate and describe subjective opinions expressed on the web or in Information Systems.
Other external ontologies that are also part of the Participation ontology are:
Simple Knowledge Organization System (SKOS): Is a W3C's standard that serves as a bridging representation model between the chaotically heterogeneous traditional data models for taxonomies and the extremely formal languages used in the Semantic Web, like the Web Ontology Language (OWL).
NLP Interchange Format (NIF): An ontology for the annotation in the NLP domain. It mainly covers the annotation of NLP analysis results through the use of nif:Context.
DBpedia: DBpedia is a community project that extracts structured, multilingual knowledge from Wikipedia and makes it freely available on the web using Semantic Web and Linked Data technologies.
