Nava, Sahar; Bunn, Roderic; Burman, Esfand; (2024) Applying Natural Language Processing to Sentiment Analysis in Building Performance Evaluation. In: Proceedings of the 2024 CIBSE Technical Symposium. Chartered Institution of Building Services Engineers (CIBSE): Cardiff, UK.

Preview

Text 22_Applying NLP to Sentiment Analysis In BPE_Final.pdf - Accepted Version Download (184kB) | Preview

Abstract

In many occupant satisfaction evaluation surveys, including the Building User Satisfaction (BUS) methodology, qualitative data is collected via text boxes to capture users' insights regarding various facets of building features. Researchers use various forms of thematic or sentiment analysis to classify or otherwise categorise written survey comments. As the subjective nature of those analyses risks the introduction of researcher bias, a more objective approach would be desirable. This study investigates the potential of employing Natural Language Processing (NLP), a subset of artificial intelligence (AI) for sentiment analysis within the context of occupant satisfaction evaluation. It examines case studies of two distinct buildings. BUS surveys, conducted at varying phases of the use stage of the buildings’ lifecycles, underwent an analytical comparison between traditional manual methods and an NLP approach. The findings showed that the AI analysis closely matched human interpretation. It is recommended for future research to use the methods of this paper for predictive analytics and provision of recommendations from textual feedback.

Download activity - last month

Export as XMLJSONCSV

Download activity - last 12 months

Export as CSVJSONXML

Downloads by country - last 12 months

Export as CSVXMLJSON

Archive Staff Only

View Item