"Do you just count words?" A summary of language analysis projects at Lancaster University

In this talk, Paul will provide a brief overview of recent and ongoing text analysis projects in the UCREL research group and other cognate centres in Lancaster University in the UK. Our research is heavily interdisciplinary, often based in real world applications which drives the research forwards and uncovers new research questions in Natural Language Processing (NLP), Corpus Linguistics (CL), Digital Humanities (DH) and Data Science (DS). Key themes include (a) spatial humanities which combines DH, NLP plus Geographical Information Systems (GIS) for analysis of the representation of place, space and time in, for example, 18-19th century literature and Holocaust survivor interviews (b) historical variant spelling detection and semantic analysis for Early English Books Online (EEBO), a Shakespeare corpus and UK and EU parliamentary transcripts (c) health communication: Metaphor in end-of-life care (MELC), Improving Peer Online Forums (iPOF) and the 4D Picture projects where we analyse online mental health and cancer forums or patient experience monitor corpora (d) Corporate Financial Information Environment (CFIE) projects working with domain experts in financial narrative analysis (e) low resource and multilingual NLP e.g. the National Corpus of Contemporary Welsh, Urdu NLP for cross lingual text reuse detection, Igbo Named Entity Recognition, and Arabic semantic analysis. Our recent key outputs and methodological contributions are in the areas of language resources, linguistically annotated corpora and open source annotation software, e.g. PyMUSAS.