mobisoc-lab projects.


a current and recent selection

DIAS_ITAUSA-UK International Technology Alliance (2016-26). The Distributed Analytics and Information Sciences International Technology Alliance (DAIS ITA) performs fundamental research on how people and computing devices can work together in rapidly changing situations, such as major disasters, to improve safety and security. The mobisoc focus concerns human groups. In situations with sparse data points, it is challenging to predict how group behaviour may evolve.  Therefore we are augmenting data-centric approaches with new generative agent-based models of groups, that can enhance predictive insights under a range of assumptions. Cardiff’s wider involvement in the ITA programme is led by Alun Preece at the Crime and Security University Research Institute.

 

 

Meeting networks (2016-21). In addition to electronic communication, a requirement of people working together in most organisations is the need to meet colleagues. Meetings provide a valuable insight into how organisations and individuals within them function, independent from the organisational structure. Applying new metrics and analysis, we investigate this using a large corpus of organisational data that reveals the extent to which meetings provide connectivity and structure within the organisation. This project is part supported by British Gas plc.

 

 

beer

Tapping into differences in alcohol cultures (2015-18). Social media enables global cultural habits and differences to be explored. Although there has been a recent surge of research looking at the reporting of food consumption on social media, the topic of alcohol consumption remains poorly investigated. One social app stands out in this regard: Untappd is an app that allows users to ‘check-in’ their consumption of beers. This allows observation of different social and drinking cultures to be explored (without direct researcher participation!). Read more on Martin and Matt’s analysis here.

 

 

tymerTymermood, mobiles & mental health (2015-18). Welsh for mood, Tymer is an Android app that allows us to examine, in new levels of detail, the human interaction and relationship with the smartphone. Sponsored by the Wellcome Trust, this project explores human well-being and smartphone usage. This involves both an in-the-wild study (launching summer 2016) and a controlled observation on a cohort of volunteers. This project is conducted in collaboration with the Neuroscience and Mental Health Research Institute.

 

 

Untitled-2mmm-e1423303768658Device cooperation for future networks – GIGAMOBILE (2015-17). Future communication systems such as 5G will have the option for short-range device-to-device communication links. These can be used to extend the range of cells, provide opportunistic networking or promote cellular network offloading. This requires cooperation where one device donates resources for another without guarantee of reciprocation. We have found fundamental rules of nature that govern this form of cooperation (forthcoming) and are further developing these for the wireless context. Our work provides insight into the evolution of human cooperation. This work is sponsored by EPSRC in collaboration with Telefonica, Keima and Samsung. Read more here.

 

homophily

Personality and homophily in online social networks (2014-17). Human personality traits govern how we tend to approach situations. We hypothesise that patterns of connectivity between similar personality types are prevalent in online human social networks such as Facebook. This builds on results concerning spatial homophily that people with similar particular personality traits have a preference for common locations, developed by assessing Foursquare user behaviour in-the-wild. Read more about this work here soon.

 

 

 

imprompduSmartphone interruptions (2013-17). Observing human behaviour and the role of the smartphone requires consideration of interruptions. This is challenging to study because asking someone if they are interruptible is itself an interruption! Therefore observational studies are required. Through a productivity app called InprompDu, evidence for multi-step decision model has been developed, from which prediction of interruptibility can be deduced. Read more about this model here.

 

 

 

interestingnessBehaviour and interestingness (2011-2015). In noisy social media it can be challenging to establish interesting content. However user response to content, for example through retweets, is a way in which interest can be detected. Instead of undertaking semantic content analysis and matching of tweets, our approach considers the human response to content, in terms of whether the content is sufficiently stimulating to get repeatedly chosen by users for forwarding (retweeting). This involves machine learning against features that are relevant to a particular user and their network, to obtain an expected level of retweeting for a user and a tweet. Read more about this model here.

 

 

Calendar-icon

Regularity in human behaviour (2009-13). Human activity is often highly predictable, based on unconscious patterns of behaviour that are driven by routine, work, family and friends. This project explores the presence and character of periodic patterns in the visits and encounters of human individuals, using a range of novel measures. Results include new decentralised techniques to detect such communities and assess their regularity.

 

 

 

logo 512x512RECOGNITION (2010-13). Funded under the EC FP7 Future Emerging Technologies programme, this multi-million pound project was led by the mobisoc-lab. This focused on understanding and emulating human abilities for recognition of important cues in the presence of noisy stimuli in online communication environments, including social media. Online and in-the-wild studies were a feature of this research. Interdisciplinary partners included psychologists, engineers and computer scientists. Read more here.

 

 

 

socnetsSOCIALNETS (2008-11). Funded under the EC FP7 Future Emerging Technologies programme, this multi-million pound project was led by the mobisoc-lab. Focusing on the increasing number of mobile and wireless devices that have the capacity to store, carry and forward data through opportunistic means, social network structures were explored as a means to create and sustain cooperative device communication. Interdisciplinary partners included physicists, anthropologists, engineers and computer scientists. Read more here.

 


Creative Commons acknowledgement.