This repository explores three main steps of text mining, namely word frequency, word cloud, topic modelling, sentiment analysis, topic-sentiment correction heatmap, bigram, word co-occurrence, speech complexity, emotional arch, dendrogram, phrases net, and sentiment stability
- text-mining part 1: word frequency analysis (figure 1: top 10 words per president), word cloud analysis (figure 2: word cloud), topic modelling - LDA analysis (figure 3: top terms per topic), NRC sentiment analysis (figure 4: sentiment distribution per president)
- text-mining part 2: topic-sentiment correlation analysis (figure 5: sentiment distribution per topic, figure 6: topic-sentiment correlation heatmap), the bigram analysis (figure 7: top 20 bigrams), network analysis of word co-occurrence (figure 8: network of word co-occurrence)
- text-mining part 3: speech complexity analysis (figure 9: lexical diversity and complexity), emotional arch analysis (figure 10: emotional arch of speech), dendrogram analysis (figure 11: presidential speech similarity), phrases net analysis (figure 12: phrases net), sentiment stability analysis (figure 13: sentiment volatility)