Bokeh provides a large set of visualisations, I have used many in my client project and they are impressed with the work, and the best part is bokeh has its own private community, so you can ask the community for help if you get stuck or if you get any problem Review collected by and hosted on G2.com.

UI can be better otherwise it is all good Review collected by and hosted on G2.com.

Bokeh in-detail elegant visualize feature providing variety to a analyst. Comparing to matplotlib , bokeh is good choice. Review collected by and hosted on G2.com.

Not default in Spyder . But Jupiter also good choice Review collected by and hosted on G2.com.

It is a powerful tool that allows users to create stunning graphics and charts for data analysis and presentation. Review collected by and hosted on G2.com.

Steep Learning Curve and very limited customization. While Bokeh provides a wide range of visualization options, it may not provide as much customization as other libraries. Users may need to rely on other tools or libraries to achieve specific visual effects. Review collected by and hosted on G2.com.

Good to learn, lots of visualisation charts to use in many mediums such as html and server. It is a good alternative to plotly. Review collected by and hosted on G2.com.

The help and resources are very limited and I am still trying to find out how I can customised the plot and theme. Review collected by and hosted on G2.com.

Helps in creating interactive graphics, data reports and can handle large datasets. Review collected by and hosted on G2.com.

Overall it is good, nothing major but needs a few more troubleshooting elements and is a little hard to debug. Review collected by and hosted on G2.com.

Lots of already supported graphs and graph types Review collected by and hosted on G2.com.

Customizing interactivity in bokeh for not-widely used plot types could be difficult Review collected by and hosted on G2.com.

Bokeh is rather easy to use, and is a fantastic new way to create visualizations! Review collected by and hosted on G2.com.

It is not the same-old-thing. While it does new/great things, it also has a learning curve. Review collected by and hosted on G2.com.

Bokeh was the first visualisation tool I tried after having a go with basic pyplot matplotlib graphics. It's super easy to get up and running and produce some beautiful interactive visualisations. I even tried my hand at getting a server up and running on AWS with great results. There's tons of tutorials out there as well as some stellar documentation. Overall I loved this tool (my boss was very impressed too). Review collected by and hosted on G2.com.

I found that hosting a Bokeh server with an Ubuntu instance wasn't exactly noob friendly. I had a few difficulties installing it and connecting to the page from the internet. It worked in the end with some troubleshooting. Review collected by and hosted on G2.com.

I like that bokeh can be easily implemented in Jupyter notebook. Bokeh documentation is also very comprehensive and user friendly. The plots that I create using Bokeh are very visually pleasing as well. I also really enjoy how Bokey provides an interactive view, a feature that matplotlib does not provide for instance. Review collected by and hosted on G2.com.

I only used Seaborn and matplotlib outside of Bokeh, and so far, all I can tell is that Bokeh is superior to these libraries in many ways. One thing to mention is that it usually requires me to write more lines of code to generate the same plot in Bokeh than the other two libraries. Review collected by and hosted on G2.com.

Interactivity, Layer-wise plot building & native ability to export. I also like that how I do not need any software and visualizations just pop open in a browser without need of a software and a erso with any desktop PC, mac or Linux can see and interact with them in exactly the same manner. Review collected by and hosted on G2.com.

Bokeh is a bit more work as compared to other interactive libraries e.g. Plotly. See my this answer for building a heatmap: https://stackoverflow.com/a/65014498/7212929. The same plot requires much more effort as compared to plotly express. I like layer-wise plot building but there should be a higher level api too which may provide all this functionality by abstraction. It could still be passed as an argument though like maybe hover = True and for details and more control current architecture may keep evolving. Review collected by and hosted on G2.com.