I have conducted sentiment analysis and topic modeling on Twitter data related to the ongoing conflict between Russia and Ukraine. Using hashtags such as '#UkraineUnderAttaсk', '#UkraineRussiaWar', '#RussiaUkraineConflict', '#RussianUkrainianWar', '#ukrainerussia', '#WorldWar3', '#RussiaVsUkraine', '#StandWithRussia', and '#StandWithUkraine', I extracted approximately 5,000 tweets.
Through sentiment analysis, I was able to determine the polarity and subjectivity of each tweet. Polarity refers to the intensity of an opinion, which can be positive or negative, while subjectivity is the degree to which an individual is personally invested in a topic, which can also be positive, negative, or neutral.
Additionally, I performed topic modeling on the collected data using the Latent Dirichlet Allocation (LDA) model. This unsupervised machine learning technique helped me to detect word and phrase patterns within the documents and cluster similar expressions into topics. The model generated eight different topics, with each topic being a combination of keywords that contribute a certain weightage to the topic.
To evaluate the model's performance, I used topic coherence, which measures the score of a single topic by assessing the degree of semantic similarity between high scoring words within the topic. This metric helped me to distinguish between semantically interpretable topics and topics that were artifacts of statistical inference.
Overall, my analysis of the Russia-Ukraine war-related tweets using sentiment analysis and topic modeling provides a deeper understanding of people's opinions and thoughts on the ongoing conflict.