ABSTRACT
The existence of coronavirus disease 19 (COVID-19) has prompted the Indonesian government to issue some policies to prevent the spread of the virus. One of them is by prohibiting homecoming (mudik) during the Eid al-Fitr holiday on 12-14 May 2021. Homecoming has become an Indonesian tradition since a long time ago. The policy raised different kinds of opinions and responses on social media. In this paper, we would like to see how people really feel about this policy. We are conducting a social media analysis focusing on the sentiment and emotion analysis about the homecoming from Twitter (as one social media platform). We used Bidirectional Long Short-Term Memory (BiLSTM) architecture as the base in modeling the classifier for both sentiment and emotion analysis. The results show an accuracy of 0.84 for the sentiment analysis model and an accuracy of 0.91 for emotion analysis. The number of positive sentiments is higher than negative sentiment and this means that there are more people who support the government policy to ban homecoming than those who oppose it. While the emotion that appears higher than the other is 'tender' and this means that Indonesians really care and love their family. © 2021 IEEE.