Check out plotly it's pretty nice and easy to get going

For the types of visualizations you're describing, the choice probably won't matter. I view matplotlib as "matlab flavor" and ggplot2 as "R flavor". For R-type work (a certain type of table-based stats) I just use R.

For matlab type work (image processing, simulations, etc) I now use matplotlib. This is mostly numpy/scipy things rather than... pandas things. Python is interesting because it has things that are beyond matplotlib (VTK, etc) and beyond matlab. Typically when you're prototyping in matlab you're assuming you will have to rewrite in a different system eventually, but with python you can move the prototype further down to more polished prototype easily.

I do a lot of image processing and am too familiar with matlab, so matplotlib generally came naturally for translating that prior knowledge. So really it depends on what sorts of things you are familiar with, languages you use, and would want to do in the future. I think with either choice you will eventually hit some wall of difficulty.

There are also more visualization and plot focused things (TeX family or PostScript and PDF) as well as the "processing" language.

I use R for... not-image-type analysis stats and generate plots in R using R's plotting. I mostly use python for matlab-type things and matplotlib seems more natural for that.

Julia is on my todo-list and I have heard good things about their plotting ecosystem but I have not looked into it.

Incidentally VTK is extremely well designed for the type of language it's based on and the problems its solving... but that's not really 2D plotting.