Exploiting user-generated content for service improvement: Case airport twitter data

Abstract

The study illustrates how airport collaborative networks can profit from the richness of data, now available due to digitalization. Using a co-creation process, where the passenger generated content is leveraged to identify possible service improvement areas. A Twitter dataset of 949497 tweets is analyzed from the four years period 2018–2021 – with the second half falling under COVID period - for 100 airports. The Latent Dirichlet Allocation (LDA) method was used for topic discovery and the lexicon-based method for sentiment analysis of the tweets. The COVID-19 related tweets reported a lower sentiment by passengers, which can be an indication of lower service level perceived. The research successfully created and tested a methodology for leveraging user-generated content for identifying possible service improvement areas in an ecosystem of services. One of the outputs of the methodology is a list of COVID-19 terms in the airport context.