As I assert in this post about displaying Likert data, my “go to” approach to any sentiment data is an “inside-out” divergent bar chart with the neutrals in a separate column, as shown [in above graphic].

If you read the very thoughtful comments you’ll see that not everyone buys into this approach. I don’t want to get into a discussion here because whether you like inside-out divergent, default sort-order divergent, or 100% stacked bar, all of them are going to run into problems when you want to show the differences in responses based on demographic breakdown.

How to visualize Likert scale data broken down by different demographics - Data Revelations