The Covid-19 virus has been one of the most discussed topics on social networks in 2020 and 2021 and has affected the classic educational paradigm, worldwide. In this research, many tweets related to the Covid-19 virus and education are considered and geo-tagged with the help of the GeoNames geographic database, which contains a large number of place names. To detect the feeling of users, sentiment analysis is performed using the RoBERTa language-based model. Finally, we obtain the trends of frequency of total, positive, and negative tweets for countries with a high number of Covid-19 confirmed cases. Investigating the results reveals a correlation between the trends of tweet frequency and the official statistic of confirmed cases for several countries.