Hello, dear Bokeh contributors and users.
I write this message in a response to a request by Brian Van de Ven, in which he asked me to clarify a series of my tweets regarding Bokeh vs. Plot.ly comparison. For the reference, the original tweet is here: https://twitter.com/boris_gorelik/status/757878933585072128. Brian has asked me to provide more details “around this statement”. Since I am not 100 sure which of my statement Brian was referring to, I will elaborate on everything I think about Bokeh, in the hope that the developing team will find this feedback useful.
-
I love Bokeh. It is currently my de-facto standard visualization library for data exploration. I use it exclusively in Jupyter notebooks. Occasionally, I share stand-alone HTML files with colleagues, and I consider using a Bokeh app for one of my future presentations.
-
Whenever I need a static image (blog post, technical report, presentation slide), I go back to Matplotlib, as I find its typography superior over Bokeh.
-
No 3D rendering is not a big deal, but sometimes I wish I could create a quck 3d interactive scatter plot.
-
The linked brushing feature was the feature that made me making my mind in my bokeh vs plotly argumentation.
-
Another “killer feature” is that in bokeh, the data is private and local by default. This is unlike plotly, in which the default behavior is to send the data to their API’s. Some of the documented features don’t work if run offline.
-
For some reason, the interactive plot.ly graphs feel more responsive, compared to @bokeh. I don’t know how to explain this – it’s just that.
I hope this helps,
Boris
