Big data reveal how people feel about rivers: Oxford geographers launch River Sentiment Dashboard

The River Sentiment Dashboard gives insight into the changing public sentiment surrounding English rivers at a time when poor river water quality and sewage pollution are making headlines. "We hope that people who care about rivers in their community will be able to use this tool to strengthen their case for improving our water environment", said Dr Helge Peters, a human geographer who led product development.

The online dashboard (www.riversentiment.app) will be launched during a public webinar on 25 May 2022 at 12pm UK time.

The team behind the dashboard applied sentiment analysis to millions of social media posts to investigate how Twitter users feel about more than 450 rivers in the Thames basin in England. The dashboard provides information about negative and positive sentiment in river-related tweets over time alongside the basic emotions and common issues that people express in river talk on Twitter.

Revealing feelings with language technology

Mining the Twitter platform for any tweets mentioning rivers in the Thames basin, the researchers used natural language processing techniques to extract affective states, such as disgust or joy, from the resulting large data set. Natural language processing enables computers to distil insights from large sets of unstructured text, such as social media posts.

The team then visualised open government data about the ecological status of rivers alongside the results of the sentiment analysis. "Natural language processing makes it possible to compare social media data about the feelings that rivers provoke with data about their environmental quality, which allows us to see the impact of the natural environment on human wellbeing," explained Dr Peters.

Alternative data for nature's value

The project developed from prior Oxford research into the valuation of blue and green infrastructure such as rivers and parks. Fieldwork discovered that planning professionals in London were frustrated with standard economic methods for appraising the value of blue and green infrastructure assets because they could not capture the benefits which local communities derive from improving river water quality.

However, members of these communities often passionately talk about the value of river water quality on social media. Finding that financial professionals use alternative data such as social media sentiment to supplement asset valuation, Dr Peters set out to put the same advanced digital tools in the hands of local communities caring for their rivers.

Designing with communities

Geographers from Oxford teamed up with catchment experts Rebecca Turnpenney and Will Oliver from the rivers charity Thames21, which annually engages 7000 volunteers to improve waterways in their communities. The team created the dashboard during a 6-month long series of design sprints, a method for prototyping innovative services together with service users.

"We engaged the rivers community throughout the design process to maximise the impact from this tool", emphasised Dr Peters. Simon Dadson, professor of hydrology and fellow of Christ Church, served as principal investigator of the project. A grant from Oxford's ESRC Impact Acceleration Award funded data science and software development by the research engineer Nathanael Sheehan.

The free and open-source River Sentiment Dashboard can be accessed at www.riversentiment.app