Abstract
Babić, Karlo Petrović, Milan Beliga, Slobodan Martinčić-Ipšić, Sanda Jarynowski, Andrzej Meštrović, Ana In this paper, we analyze and compare Croatian and Polish Twitter datasets. After collecting tweets related to COVID-19 in the period from 20.01.2020 until 01.07.2020, we automatically annotated positive, negative, and neutral tweets with a simple method, and then used a classifier to annotate the dataset again. To interpret the data, the total number as well as the number of positive and negative tweets are plotted through time for Croatian and Polish tweets. The positive/negative fluctuations in the visualizations are explained in the context of certain events, such as the lockdowns, Easter, and parliamentary elections. In the last step, we analyze tokens by extracting the most frequently occurring tokens in positive or negative tweets and calculating the positive to negative (and reverse) ratios.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
