Topic modeling on AirBnB reviews during the pandemic

Abstract

In the current dissertation, we are going to exploit LDA topic models to analyze AirBnB online comments and discover meaningful patterns. Topic modelling is a well-known and prevalent tool to extract concepts of small or large text corpora. These text collections often enclose hidden meta groups. Valuable information on online reviews is often ignored, therefore our study will concentrate on extracting important and profitable business insights. Moreover, this research project aims to provide a clear understanding of how COVID-19 pandemic has influenced the tourism industry and how AirBnB has dealt with this unique and unfamiliar phenomenon. To be more precise, this analysis consists of gathering data from Get the Data - Inside AirBnB. Adding data to the debate. We will handle data from spring of 2020, which was the initial period that CoVID-19 affected Greece. Before applying LDA algorithm, we are going to implement preprocessing techniques. Preprocessing is the process of bringing your text into a form that is predictable and analyzable for your task, fitting it to a certain schema. After that, we will implement and train our model so that we obtain the results that will be evaluated.