Skip to main content

School of Social and Political Science: Staff profiles

Search

Clare Llewellyn

Clare Llewellyn
Name
Dr Clare Llewellyn
Title
Career Development Fellow
Address
2.03, 2F2 18 Buccleuch Place Edinburgh UK EH8 9LN
Telephone
+ 44 (0)131 650 6633
Email
Research Interests
Neuropolitics of identity, Social media, Big data, Natural language processing, Text and data analytics
URL
http://www.sps.ed.ac.uk/staff/politics/clare_llewellyn

Guidance and Feedback Hours

  • I am generally available Tuesday - Friday, please email to make an appointment

I am interested in thedevelopment and definition of cross-disciplinary methodological and ethical techniques and standards in the GovTech domain. I have extensive experience in working in a collaborative trans-disciplinary environment and have produced co-authored papers with academics from informatics, politics, social work, history, botany, criminology, linguistics, information studies and library science. More recently, I have specialised in collaborating with social scientist and I have been embedded in a trans-disciplinary lab where I have expanded and exploited my unique combination of skills. 

Specifically, I have developed novel data analysis techniques, underpinned by data science technologies such as natural language processing, supervised and unsupervised machine learning and statistical analysis. My recent focus has been in developing and implementing studies to gather and analyse social media data and in using these results either in combination with, or as an underpinning for, experimental and quantitative social science research. My work in this area is unique as I use a longitudinal multiple aspect collection approach to gathering social media data. This involves gathering several parallel datasets over extended time periods allowing the observation of changes, the ability to map to changes in data to changes in public opinion and enables analysis from multiple viewpoints.

I have accumulated what is probably the largest corpus of Brexit-related tweets currently available. The scale of this datapresents significant storage, processing and visualisation issues that I respond to with innovative methodological solutions. I have used this data to answer political questions that give a real-world impact. These questions have included: Can we determine if it was social media that popularise the Vote Leave campaign message that we should “take back control of £350 Million to spend on the NHS”? Were Russian Twitter Trolls designed to influence the US 2016 Presidential Elections active in the Brexit debate? Where did the supporters of Tommy Robinson, the former English Defence League leader claim to be located? Was the 2017 General Election Twitter conversation dominated by discussion of politicians or substantial issues?