Development of a Voice Virtual Assistant for the Geospatial Data Visualization Application on the Web

Abstract

Voice assistants can elevate interaction in geospatial data web platforms. This research introduces a voice assistant in the BStreams platform and focuses on understanding user commands in the geospatial domain. We developed a specialised geospatial discourse framework through structured prototyping. A survey with 66 participants revealed prevalent English geospatial terminologies. Using ChatGPT, we found the term suggestions aligned with survey results, with a notable correlation (r = 0.81, p < 0.01) between the NPL model’s probability scores and term prevalence in survey data. Our study also incorporated usability tests on the application, which uses tools like Web Speech API, Leaflet, and Mapbox geocoding. Results from these tests reaffirm the potential of voice assistants in enhancing geospatial data visualisation, though challenges persist in areas like language understanding and domain knowledge. The paper advocates for further research to refine the integration of voice technology in this domain.