Chapter 3 Urban Water Sentiment Map
We demonstrate the prototype of an interactive web mapping application which displays information about 60 water bodies and wetlands in the London management catchment. Users can access four main kinds of environmental information about each water body:
- its current ecological status,
- the social agencies and
- material agencies which together prevent the water body from attaining good ecological status, and
- the sentiment of tweets mentioning the water body by name.
In addition to exploring information about local water bodies, users are invited to contribute their own sentiment to the map by authoring tweets.
The map performs the ontological multiplicity and socio-material hybridity of urban water by contrasting user sentiment about water bodies as places with Defra data on material issues preventing waters from reaching good status and the sectors of society contributing to them. As a result, water bodies are mapped simultaneously as objects enacted through scientific measurement and expert classification and through place-based practices of everyday environmental engagement.
Rather than representing public opinion, the prototype aims to function as an interactive device for gathering an issue public that allows users to actively participate in social scientific knowledge production about an environmental knowledge controversy. Our previous two projects have demonstrated the capacity of computational methods to identify and describe the formation of digital water publics. However, these projects exploited the metric affordances of a social media platform in order to produce social scientific knowledge about the way in which users participate in the formation of a public. Whereas public participation in environmental controversy forms the object of such research, it remains highly asymmetrical insofar as participating publics are treated as passive providers of data (Marres, 2017). In contrast, this project deploys the interactive affordance of Twitter in order to develop a public and participatory way of knowing digital water publics.
The information displayed had been selected based on the issues which publics were likely to engage with. The previous two projects identified that users were likely to engage with questions of environmental quality, had a keen sense of place, and responded strongly to questions of responsibility. As a result, information on ecological status and the social and material agencies preventing water bodies from reaching good status is based on data provided by Defra. The data underlying sentiment analysis are mined from Twitter on a weekly basis with a search function containing a list of names of water bodies in the London management catchment, which is taken from Defra but adapted to everyday language.
The project resulted from an exploration of the relationship between the sentiment of tweets about blue and/or green spaces in London and other geospatial data. Whereas we found local instances where ecological status and index of multiple deprivation correlate, the data set of tweets was too small to produce any discernible patterns. Without additional resources, access to historical tweets and green space data remains a barrier. However, continued mining will increase the tweet sample over time. The relative paucity of geotagged tweets as well as the ambiguity following from geolocation based on natural language/place names both provide further challenges.
The prototype can be accessed under the following link