Leibniz's Monadology mentions perceptional and sentimental variations of the individual in the city. It is the interaction of people with people and events. Film festivals are highly sentimental events of multicultural cities. Each movie has a different sentimental effect and the interactions with the movies have reflections that can be observed on social media. This analysis aims to apply distant reading on Berlinale tweets collected during the festival. On contrary to close reading, distant reading let authors to observe patterns in large collection of data. The analysis is temporal and sentimental in multilingual domain and strongly positive and negative time intervals are analysed. For this purpose, we trained a deep sentiment network with multilingual embeddings. These multilingual embeddings are aligned in latent space. We trained the network with a multilingual dataset in three languages English, German and Spanish. The trained algorithm has a 0.78 test score and applied on Tweets with Berlinale hashtag during the festival. Although the sentimental analysis does not reflect the award-winning films, we observe weekly routine on the relationship between sentimentality, which can mislead a close reading analysis. We have also remarks on popularity of the director or actors.