Line chart: Compares the trend of stock prices and sentiments for the given period of time. For the stock price the daily value considered is the closing price of the day. Prices are normalized by setting the maximum to 1 and scaling down the others. For sentiment the degree of positivity for each day is calculated as follows: 1 * (positive tweets)% + 0.5 * (neutral tweets)% + 0 * (negative tweets)%. The obtained values are normalized. Hovering on each datapoint reveals the actual rates for both datasets.
Correlation coefficient: Shows the statistical relationship between the stock prices and positivity, regardless of the day of the event.
Doughnut chart: Shows the aggregated sentiment for the whole world as a percentage.
World map: Shows the aggregated sentiment per continent as a percentage.
Cluster chart: Clustering of the topics is visualized in two dimensions with the size of the clustered topics depicting the weight of a particular topic in the analyzed event.
Bar chart: Shows the word frequencies per topic and occurrence in all of the tweets of this event. The relevance score2 is automatically set at 1. With this setting the sole relevance is based on the probability of a certain term "w" from the vocabulary to find itself in topic "k". An optimal value of "λ" or the relevance score is 0.6. Moving it slightly above or below this threshold will serve for the best interpretability of the topics.