# Unsupervised Machine Learning | # Latent Dirichlet Allocation | # Text Analytics
The interactive pyLDAVis charts as a result of topic modelling conducted over five journals is put up below. Each chart consist of two panels – the left and the right. The left panel visualises the topics as circles in the two-dimensional plain by computing the distance between the topics (Sievert and Shirley 2014), (Chuang, et al. 2012) . The right panel consists of a pair of overlaid bars representing topic-specific frequency of the term as well of corpus-wide frequency of a given term. Each bubble on the Intertopic Distance Map represents a topic. The larger the bubble, the more dominant is the topic. Big non-overlapping bubbles scattered thoughout the Intertopic Distance Map is indicative of a good topic model. The words and bar updates as the cursor is moved over the bubbles or the topic is selected by the input/buttons provided. The words selected are the salient keyword that describe the selected topic/cluster. Further, on hovering over any of the salient word in the right plot, the clusters/bubbles can be backtracked. The lamda (λ) slider allows to rank the terms according to term relevance. Moving the slider allows to adjust the rank of terms based on much “relevant” they are for the specific topic. This can alter the ranking of terms which can aid topic intepretability.