ABSTRACT
Sentiment Analysis is an emerging research area that focuses on
extracting semantic and emotional inferences from natural language,
paving the way for analyses that deal with a high volume of
textual data. The growing importance of data in strategic decisionmaking
and the recognition of social networks as vast repositories
of public opinion have propelled this study, which aimed to explore
the interaction between human emotions and motorsport events.
Thus, this study focused on applying Natural Language Processing
to extract and analyze sentiments expressed in tweets about Formula
1. Advanced machine learning and deep learning techniques
were employed to train various models in the sentiment classification
task. Among these, Logistic Regression and LSTMs stood out,
achieving accuracies of 78.21% and 78.08%, respectively. The LSTM
model, in particular, was implemented on a public dataset of tweets
collected during the 2021 and 2022 Formula 1 seasons. The model
was used to classify the sentiments expressed by fans, allowing
for an exploratory analysis of data correlated to specific events of
the races. The findings revealed significant engagement patterns,
with notable spikes in emotional reactions coinciding with critical
moments of the seasons. These discoveries illustrate how particular
events can profoundly influence the emotions and behavior of fans.
From a detailed analysis of expressed sentiments, valuable data can
be obtained that may be leveraged for developing more effective
marketing and communication strategies in the sport.