Access to credit and fintech: A lexicon-based sentiment analysis application on Twitter data

Abstract

This study examines how the interaction between access to credit and new fintech ventures evolved during and after the COVID-19 pandemic by conducting three different lexicon-based sentiment analyses using NLTK, TextBlob, and Flair Python libraries. We previously gathered data from Twitter (subsequently rebranded as X) by applying different combinations of keywords in our scraper script to better understand the phenomenon and enhance the quality of the final dataset. We defined the most appropriate set of keywords that we subsequently used for analysis. We also empirically estimated whether the results obtained could be generalized to the continents involved. Although the keywords “access to credit” and “fintech” show a slight decrease in tweets at the end of the COVID-19 pandemic, we obtain meaningful insights at the continent level concerning variations in sentiment over the analyzed period. Furthermore, the most recurrent keywords show significant correlations.