Abstract

The £10.2M NERC-funded GALLANT (Glasgow as a Living Lab Accelerating Novel Transformation) project is acquiring and producing data on Glasgow's ‘living space’, for effective environmental and ecological planning and management. The key to a thriving city transition is an understanding of citizens' priorities regarding environmental and societal issues. This can be informed by data and will increase engagement in the local communities. However, such data are various and complex potentially coming from structured and unstructured sources and of quantitative and qualitative forms. Examples include deprivation indices, interview transcripts, and photos and text from social media posts, for which information extraction is not straightforward. Within GALLANT the data and analytics team are exploring novel deep learning pipelines combining natural language processing and image processing to enable information extraction from these unstructured data. Such analyses would also allow integration with qualitative data obtained through social media for example, with official statistics (e.g., SIMD). The power of this approach lies in the potential additional information obtained by processing and analysing both images and text (i.e., multimodal data) generated by the users. These pipelines will enable integration of quantitative and qualitative data potentially providing additional information to investigate the interconnections between environmental and societal patterns in Glasgow.