Ewch i’r prif gynnwys

SENTINEL (OSCAR)

Mae'r cynnwys hwn ar gael yn Saesneg yn unig.

SENTINEL: localised situational awareness via social media

The Sentinel platform provides a multi-level interface into information extracted from the social Web. The purpose of this platform is to evaluate to what degree social media data can be converted into actionable intelligence relating to public-interest events and topics in terms of reliability, usability and timeliness. We focus upon five key questions that can allow human analysts to draw conclusions as to:

  • What is happening?
  • When is it taking place?
  • Where is it occurring?
  • Who are the sides involved?
  • Why is this happening and why does it require response?

Users of Sentinel are able to direct data collection towards specific key words and geographical locations creating “channels” of interest by which intelligence can be derived.

Data driven information, knowledge driven identification

The data driven text mining tool FlexiTerm has been incorporated into the interface, allowing for a near real time summary of prevailing terms currently found within a channel, with the output being presented spatially and temporally.

Running parallel to topic identification is the indexing of data against an expertly curated ontology of criminal and extremist topics. This provides users with the ability to delve deep into their data channels to identify and track sensitive narratives.

A conversational natural language query facility called Moira (Mobile Information Reporting App) provides Siri-style assistance to users in performing tasks such as profiling individual social media posters.

Sentinel has been designed as an open platform that allows social and computing scientists to co-design useful analytic components and apps, able to help add metadata and meaning to social media data. Sentinel has been tested in a number of pilot studies including:

  • Scanning of social media traffic in relation to geographic regions including a major city and a medium- density city region.
  • A longitudinal study of a high-profile crime and its effects over a ten-month period from perpetration to sentencing.
  • A real-time study of a major planned event in a city region, including the build-up over a three-month period, and immediate
    aftermath.

Dealing with 'Big Data'

We take advantage of the pervasiveness of Cloud Computing to create a dynamic and scalable approach to data processing that provides the ability to extract near real time meaning from large volumes of data.

This takes the form of a data processing pipeline in which fragments of social media text are filtered, transformed and enriched by a pool of processing and analytic tools.

References

  1. Spasić, I., Greenwood, M., Preece, A., Francis, N. and Elwyn, G. (2013). FlexiTerm: A flexible term recognition method. Journal of Biomedical Semantics 4(27)

  2. Preece, A., Braines, D., Pizzocaro, D. and Parizas, C. (2014), Human-machine conversations to support multi-agency missions, ACM SIGMOBILE Mobile Computing and Communications Review, 18

Funding acknowledgement

The research that enabled development of the Sentinel platform was funded by the European Commission under the project "Tackling Radicalisation in Dispersed Societies (TaRDiS)", and the ESRC via the project "After Woolwich: Social Reactions on Social Media" (ES/L008181/1). Enabling research for the Moira tool was sponsored by the US Army Research Laboratory and the UK Ministry of Defence and was accomplished under Agreement Number W911NF-06-3-0001. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the US Army Research Laboratory, the US Government, the UK Ministry of Defence or the UK Government. The US and UK Governments are authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation hereon.