Example PhD

Crowd sourcing Data for Feeding Probabilistic Risk Models

Supervisor: Dr P. Burnap

Keywords: Risk, internet of things, crowd sensing, probablistic modelling

Many approaches to organizational risk assessment now follow a “tick box” mentality and are based on qualitative (e.g. harmless, dangerous, catastrophic) or ordinal (e.g. low, medium, high) scales without any accurate statistical data. This can lead to unfounded risk management strategies where risks are over/underestimated with disastrous effects. The aim of this project is to investigate probabilistic risk models that allow likelihood and impacts of potential risks to be calculated, using statistical data, and to gather data from crowd sourced Web resources to feed the model. For example, geographic data (e.g. weather prediction and seismic activity data), financial market data and traffic flow data could all be possible risks to organizational operations. There are several Web-based resources that provide this data as a live “feed”. Data may also be scraped from social media feeds, news feeds and blogs (via RSS) and transformed to feed a probabilistic model (e.g. rise of tension in specific countries, political and social support for tighter industry-wide regulation). Open Data could also be used (e.g. recent crime reports).

Key Skills/Background: Programming skills and general understanding of Internet computing. The rest can be learned as part of the PhD.

Contact: Dr P. Burnap to discuss this research topic.