Professor Paul Baker (Lancaster) delivered the 2019 Sinclair Lecture at the University of Birmingham on 24 June.

Human beings currently create around 2.5 quintillion bytes of data every day. The ability to quickly and accurately identify trends and linguistic patterns across massive and continuously growing data-sets is viewed as important although automated techniques are not yet able to out-perform human analysts in numerous ways.

In this talk I focus on what a corpus linguistics approach can offer in an increasingly crowded computational analysis field by describing an ESRC-funded project which explored how corpus methods could be used to improve NHS services. The project involved the analysis of 200,000 pieces of feedback left by members of the public on the NHS Choices website and resulted in several challenges relating to aspects of the data and the research questions that were given to us by the Patients and Information Directorate at NHS England.

A key feature of this involved a learning curve where we realised that our initial understandings around language use in this context and the right techniques of analysis were not always as precise as they could have been. As we refined our methods we found that some of the most interesting aspects of the analysis resulted in answers to questions that we had not asked. However, the process of engaging with this unfamiliar form of data and being set questions we would not have chosen ourselves ultimately resulted in benefits - we were able not only to gain insight into the nature of the data, but insight into how corpus approaches can be more effectively used for close analysis of large datasets.

The talk concludes with a discussion of the extent to which insight alone is the key to making impact.
[ Ссылка ]

свернуть