Easy to learn and use, good for basic interactive charts. Allows you to provide charts in many mediums (html, notebook and server). Good alternative to plotly and pygal.

I like that it's fairly easy to create dynamic html visualizations that look slick and feel good. Since I learned R before python for statistics and visualizations, I definitely prefer R's ggplot2 syntax (which plotly can then easily convert to an html version with plotly::ggplotly()). However, for the python work that I do (when my coworkers prefer python notebooks, etc.) the capability of bokeh is great! The api is fairly consistent across different types of plots which is great.

The library has a lot of potential to create a rainbow of visualizations. I like that the dashboards are interactive.