Home > Research > Publications & Outputs > Fuzzy tagging and processing of semantic vaguen...
View graph of relations

Fuzzy tagging and processing of semantic vagueness for crowd-sourcing public perceptions of environmental change

Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSNConference contribution/Paperpeer-review

Published
  • Fernando Sanchez-Trigueros
  • Stephen Carver
  • Jonny Huck
  • Roian Matt
  • Brooke McBride
  • Linda Moon Stumpff
  • Alan Watson
Close
Publication date05/2014
Host publicationDigital conservation: understanding the impacts of digital technology on nature conservation
<mark>Original language</mark>English

Abstract

In this presentation we illustrate the application of fuzzy tagging and a "spray and say" survey method for the documentation of community-based perceptions of environmental processes. Spatial data collection by fuzzy tagging takes account of geospatial vagueness on locating phenomena, so it moves from point and polygon abstractions through the "spraying" of places towards heat maps of spray incidence. Additionally, the "spray and say" mechanism allows the interviewee to add comments about tagged places to add context and qualifications to these spatial patterns. The grid-based Tagger (Waters and Evans, 2003; Evans and Waters, 2007) and its Map-Me PPGIS vector-adaption (Huck and Carver, 2012) are pioneering instances of both approaches.

Post-processing of "spray and say" data typically involves the treatment of the spatiotemporal dimension, text data and measures of processes co-located with tagged sites (e.g. terrain models or pollution indexes). To proceed with an integral analysis by which all such dimensions can be jointly taken into account, we combine Artificial Intelligence techniques, Natural Language Processing (NLP) and data mining for the coupling of GIS and linguistic functionality in Information Systems capable of storing, managing, retrieving, visualizing and analyzing complex multimodal data.

We present an application of the Map-Me tool, of a multimodal Information System and of an NLP-based query language to a case study on public perceptions of environmental change on the Flathead Indian Reservation (Montana), focusing on the effects of fire suppression in fire-adapted ecosystems and on the restoration of prescribed fire as perceived by tribal and non-tribal residents. The demographic, semantic and spatia properties of responses show different tagging behaviours between tribal and non-tribal residents that positively perceive the potential reintroduction of prescribed fire, depicting opposing patterns with regard to the geographical subregions and land-use categories that these cohorts tend to tag. We conclude with suggestions as to how this approach might be applied across a range of conservation projects to both ground truth and validate traditional models and survey techniques.