Research by a group of Lancaster University researchers has shown that social media sites such as twitter can detect drinking patterns in the UK including spotting the variations on national holidays and celebrations.
The researchers (PhD student Daniel Kershaw from the HighWire Doctoral Training Centre, Lecturer Dr Matthew Rowe from the School of Computing and Communications and Dr Patrick Stacey from Lancaster University Management School) who wrote the paper "Towards Tracking and Analysing Regional Alcohol Consumption Patterns in the UK through the use of Social Media? were recently awarded Best Student Paper at the international WebSci14 conference which was held from 23 to June 26, 2014.
Ever-increasing drinking levels have led to calls from public services (e.g. police and health services) to assess the effect it is having on people and society.
Current research methods that monitor rates of alcohol consumption across the UK are costly, time consuming, and do not supply sufficiently detailed results as they look at snapshots of individuals' drinking patterns, which rely on generalised usage patterns, and post consumption recall.
As an alternative, this new research looks at using the much cheaper and faster tools of social media, such as Twitter, to monitor the rate of alcohol consumption in regions across the UK by introducing the Social Media Alcohol Index (SMAI).
By looking at the variation in term usage, and treating the social network, which also records location and time, as a self-reporting sense-network, the SMAI will show variation in drinking patterns on both local and national levels within the UK.
The study used 31.6 million tweets collected over a 6 week period, and used the Health & Social Care Information Centre (HSCIC) weekly alcohol consumption pattern as a ground truth.
High correlations between the ground truth and the computed SMAI (Social Media Alcohol Index) were found on a national and local level, along with the ability to detect variation in consumption on National holidays and celebrations at both local and national levels.
Daniel Kershaw said, "Taking this forward we are looking into the variations of the language used across different geographical locations and social groups. In the near future we will be aiming to enhance our work by collaborating with local public services in order to gain feedback on real-world applications of this research."
On the graph below the green line is the HSCIC data, blue is the scores, and red is the moving average.