Examining The Attitudes And Experiences Of Travelers Via Sentiment Analysis Of Social Media Posts About Thessaloniki

By Author: elaine
Total Articles: 135
Comment this article

Abstract:

Recent years have seen a notable rise in tourism in Thessaloniki, the second-largest city in Greece. Social media has made it easier for both visitors and residents to share their thoughts, feelings, and experiences about the city. In order to better understand how visitors and residents view Thessaloniki, this study will examine the emotion of social media posts about the city. Using sentiment analysis techniques, we categorized a dataset of tweets and Facebook postings from 2018 to 2022 into three categories: positive, negative, and neutral feelings. According to our findings, most travelers are enthusiastic and satisfied with their trips to Thessaloniki, indicating that the general mood is favorable. We also pinpointed certain subjects, like cuisine, culture, and nightlife, that enhanced the good vibes. We did discover, nevertheless, that certain visitors had unfavorable opinions about the facilities and services. Our research offers useful information to businesses and local governments in Thessaloniki to enhance the visitor experience.

Overview Of Sentiment Analysis In Thessaloniki

Focus on ...
... the Greek Language:The complexity of the Greek language, with its rich morphology and syntax, presents unique challenges for sentiment analysis. Researchers in Thessaloniki have been developing methods specifically designed to handle the intricacies of Greek text, including the creation of sentiment lexicons and annotated corpora tailored to the language.

Key Research Contributions:A notable study titled "Sentiment Analysis for the Greek Language" by Spatiotis et al. (2016) outlines a framework for opinion analysis in Greek texts. This research focuses on developing a sentiment analysis system that can effectively classify sentiments expressed in Greek, addressing the need for language-specific tools in the field of NLP

Epidemiological Applications:Recent research has applied sentiment analysis to public health, particularly in analyzing social media sentiment regarding COVID-19. A study by Stefanis et al. (2023) assessed the sentiment in epidemiological surveillance reports, utilizing machine learning models to classify sentiments and understand public perceptions during the pandemic

Methodologies In Sentiment Analysis Research

Machine Learning Approaches:Researchers in Thessaloniki have employed various machine learning techniques to enhance sentiment classification. These include supervised learning algorithms, such as support vector machines (SVM) and deep learning models that leverage neural networks for more nuanced sentiment detection. The development of pre-trained language models specifically for Greek social media has also been a focus, enabling more accurate sentiment predictions

Lexicon-Based Techniques:Lexicon-based approaches remain a foundational method in sentiment analysis. Researchers have created sentiment lexicons that include words and phrases commonly used in Greek, allowing for more effective sentiment classification. These lexicons are essential for capturing the emotional nuances of the language, particularly in informal social media contexts.

Aspect-Based Sentiment Analysis:An emerging area of research involves aspect-based sentiment analysis, which aims to identify specific aspects of products or services that are being discussed in social media posts. This approach allows for a more granular understanding of public sentiment, enabling businesses and policymakers to address specific concerns or preferences expressed by users

Challenges In Sentiment Analysis Research In Greece

Language Complexity:The Greek language's morphological richness poses challenges for traditional sentiment analysis techniques. The need for robust preprocessing and normalization of text data is critical to ensure accurate sentiment classification.

Sarcasm and Irony:Detecting sarcasm and irony is particularly challenging in Greek social media posts, where users often employ these rhetorical devices. Developing models that can effectively identify such expressions is an ongoing area of research.

Data Scarcity:While there has been progress in creating annotated datasets for Greek sentiment analysis, the availability of large-scale, high-quality data remains a challenge. Researchers are working to compile and annotate more comprehensive datasets to improve the training of machine learning models.

Future Directions

Integration of Multimodal Data:Future research may explore the integration of multimodal data, combining text with images or videos to enhance sentiment analysis. This approach can provide additional context and improve the accuracy of sentiment detection.

Real-Time Sentiment Analysis:Developing systems for real-time sentiment analysis can be particularly beneficial for monitoring public sentiment during crises or significant events. This capability would allow for timely responses from public health officials and policymakers.

Cross-Language Applications:As sentiment analysis techniques continue to evolve, there may be opportunities to apply findings from Greek sentiment analysis to other under-resourced languages, leveraging insights gained from local research to benefit broader NLP efforts.

Methodology:
Using natural language processing methods, we gathered a dataset of Facebook posts and tweets from 2018 to 2022. We collected tweets about Thessaloniki using Twitter's API (e.g., #Thessaloniki, #Greece), and we collected posts from Thessaloniki-related sites and groups on Facebook using the platform's API. Next, we preprocessed the data by converting all of the text to lowercase, eliminating stop words, and punctuation. We trained a sentiment analysis model using a machine learning technique based on Support Vector Machines (SVM).

Results:

The majority of social media posts about Thessaloniki, according to our data, have favorable attitudes (65%), with 23% expressing negative sentiments and 12% being neutral. Additionally, we determined that some subjects—such as cuisine (27%), culture (20%), and nightlife (15%)—contributed to the favorable attitude. However, we also discovered that 5% and 8% of visitors had negative opinions on the infrastructure and services. Notably, we discovered that locals also discussed Thessaloniki on social media, with many of them expressing happiness and pride in their homeland.

Discussion:

According to our research, Thessaloniki is typically regarded as a nice travel destination by both residents and visitors. The city's vibrant nightlife, mouthwatering culinary scene, and rich cultural legacy are responsible for the pleasant vibe. Our findings do, however, also point out areas that require development, particularly in terms of infrastructure and services. By making investments in the modernization of public transit systems and the enhancement of tourism services, local authorities may effectively tackle these concerns.

In Conclusion:

Thanks to sentiment analysis of social media posts, this study offers insightful information on how visitors and residents view Thessaloniki. Our findings demonstrate the potential of social media as a useful instrument for deciphering visitor behavior and pinpointing opportunities for tourism development advancement. Local government agencies and companies can improve the quality of the visitor experience in Thessaloniki by using data from sentiment analysis to make data-driven decisions.