A seminar in the Bristol Data Science Seminar Series hosted by the Jean Golding Institute and the Heilbronn Institute, University of Bristol

Many healthcare providers within the National Health Service (NHS) routinely collect patient feedback to evaluate service quality from a patient's perspective. It is common to ask patients free text questions to describe ‘what was good’ and ‘what could be improved’. Text Mining (e.g. Text Classification and Sentiment Analysis) coupled with interactive dashboards can help process large volumes of feedback efficiently and convert them into actionable insights. This also allows us to monitor the effectiveness of any changes in service feedback over time. We are currently developing such a tool for the Nottinghamshire Healthcare NHS Foundation Trust, hoping that it will be adapted across the NHS in the future.

During our presentation we will describe two key ideas that we try to follow in our work: (i) working openly by open-sourcing the analysis code and data where possible to promote replication, reproducibility, and further developments; and (ii) automating common steps in our workflow by developing R/Python packages to increase usability and documentation. To showcase these techniques, we will demonstrate how we implement these ideas by focusing on two parts of our project: (i) working with databases in R; and (ii) developing dashboards with Shiny and Golem.

Register for your free place

About the speakers: Andreas Soteriades: I am an Applied Mathematician with Data Science experience in both academic and non-research areas, from dairy farm efficiency modelling to Text Mining (plus a few other in between!). I am currently crunching NHS patient feedback data with Scikit-learn (Python) and tidytext (R) and translating findings into an illuminating dashboard (Shiny/Golem) for managers and non-technical staff. I have been an R user for a long time now but have recently discovered the world of Python with much enthusiasm!

Milan Wiedemann: I’m a data scientist working for the Clinical Development Unit at the Nottinghamshire Healthcare NHS Foundation Trust. I enjoy working on software tools in R that help to visualise and analyse data. My main research interest is evidence-based health care, particularly in clinical psychology. During my PhD in Experimental Psychology I tried to understand more about how psychological therapies for anxiety disorders work and also developed an interested in data science.