Designing an infographic

Topic modeling algorithms identify distinct topics through their mathematical operations. Now, the question is how to make the results easily interpretable.

The numerical results of topics are difficult for humans to understand. Humans are not good at processing large numbers of vectors and matrices or deriving insights from an ocean of numbers. Users may lose interest after inspecting a lot of dry numbers. For example, a user who is not a data expert may ask what it means when a word is multiplied by a number, like the following:

'0.017*"said" + 0.016*"deal" + 0.013*"agre" + 0.012*"million" + ' '0.012*"compani" + 0.011*"union" + 0.011*"european" + 0.010*"agreement" + ' '0.010*"billion" + 0.008*"contract"

We need a better visual tool!

But in our defense, the topics discovered by models mechanically may be truly nonsense. The...