Κατεύθυνση Ψηφιακά Μέσα Επικοινωνίας και Περιβάλλοντα Αλληλεπίδρασης
Library of the Faculties of Political Science and Public Administration, Communication and Mass Media Studies, Turkish and Modern Asian Studies, Sociology

2017-12-15

2017

Bairaktaris Ioannis

Κωνσταντίνος Μουρλάς. Αναπληρωτής καθηγητής. Τμήμα Επικοινωνίας και Μέσων Μαζικής Ενημέρωσης. ΕΚΠΑ

Ανάλυση Συναισθήματος στο Twitter για τους δύο Έλληνες πολιτικούς αρχηγούς, Αλέξη Τσίπρα και Κυριάκο Μητσοτάκη

Greek

Twitter Sentiment Analysis: Greek Political Leaders, Alexis Tsipras & Kyriakos Mitsotakis

Social media are now an integral part of everyday life. The information that exists in these social media is huge and is constantly increasing. Therefore, it is logical to have the need to collect, process and analyze these data, so they can help on decision making. Such work is done by sentiment analysis and one of the most important areas that is applied is that of politics.
This thesis focuses on the popular social networking tool, Twitter. Political leaders, parties and politics in general are among the most commented topics by Twitter users, so we chose this area to apply sentiment analysis. In particular, we dealt with the two Greek political leaders, Alexis Tsipras and Kyriakos Mitsotakis, and tried to find out what the people’s sentiment about them (negative, positive, neutral) is, based on twitter comments.
Sentiment analysis was performed using «R programming language» on the «R Studio» platform. We collected a total of about 13.500 tweets referring to the two Greek political leaders using keywords, in a period from January 2017 to June 2017. We then implemented an algorithm that «cleaned» the tweets to remove the useless information and finally we applied sentiment analysis to tweets using two sentiment lexicons (positive and negative) that we created. To evaluate the algorithm, we picked up some of the tweets we collected and manually annotated them with the help of three annotators. By comparing the results we concluded that the algorithm we implemented works to a satisfactory degree.

Social, Political and Economic sciences

Sentiment Analysis, Twitter, Politics, R programming, sentiment lexicon

Yes

2

Yes

84

102

File access is restricted only to the intranet of UoA.

https://pergamos.lib.uoa.gr/uoa/dl/object/2399353