About
The NIHR Innovation Observatory’s Soft Intelligence Squad (NIHRIO SIS) are monitoring, tracking, and analysing ‘soft intelligence’ to explore health impacts.
Mobilised in March 2020, the group are using novel text mining and machine learning techniques, involving sentiment detection, analysis, and classification, to help extract and synthesise meaningful insights from the ‘public voice’.
Led and coordinated by Dr Christopher Marshall, members of the NIHRIO SIS include:
- Dr Georgie Wilkins
- Katherine Lanyi
- Rhiannon Green
- Savitri Pandey
Currently, the group are focusing their efforts on identifying, collating, and synthesising soft intelligence to explore the impact of the Covid-19 pandemic and resulting lockdown measures on wider health issues. Detail on some of the work the group are doing is described below.
If you would like more information about the group and how they might support your work, please contact chris.marshall@io.nihr.ac.uk
Project: Exploring the impact of Covid-19 on mental health: Longitudinal study of UK tweets
Covid-19 is having a profound impact on people’s mental health. As we move through the pandemic, there have been calls to try and find useful ways to survey, collect and analyse the long-term impacts on mental health across the population.
Studies have shown that twitter can be a useful and valuable resource for obtaining real-time health data using crowdsourcing methods. In particular, twitter has been used in the past to track trends and disseminate health information during previous viral pandemics, including H1N1, Ebola, MERS and the Zika virus.
Since April 30, NIHRIO SIS researchers have been prospectively scanning for and tracking UK-based tweets relating to Covid-19, the lockdown and mental health. More specifically, the squad are exploring and analysing tweets on the following:
- Overall mental health
- Anxiety
- Depression
- Stress
- Suicide
The overall volume and sentiment of tweets relating to these issues are updated daily. The live, interactive charts below visualise the flow of volume and sentiment over time relating to ‘overall mental health’ on page 1 and ‘anxiety’, ‘depression’, ‘stress’ and ‘suicide’ on page 2.
Further, the group are tracking and examining the underlying topics, trends and patterns emerging from the data on a weekly basis, through topic clustering and qualitative coding. The group are working on a way to usefully visualise these findings, which will be shared here soon.