arclight on Nostr: Python people: is there a decent publication quality plotting library that can make ...
Python people: is there a decent publication quality plotting library that can make basic black and white plots, write them in a common format (PNG, SVG), and does not depend on matplotlib?
matplotlib is brittle, unintuitive, and basically unusable. I do not want or need to display plots, I do not want or need web or Jupyter support or any interactivity.
I need something that runs locally and just turns data into lines and points and text in a graphics format that can eventually be printed. Something like what gnuplot does but that can be driven from Python.
Everything either is aimed at "data visualization" on the web or is just a layer on top of matolotlib. I have two columns of numbers, I want a professional-looking B/W scatter or line plot that can be incorporated into a press-friendly document.
plotnine is failing due to some weird-ass matplotlib backend issue that the plotnine maintainer claims is because something is installed wrong. I've spent my alotted hour trying to get the minimum functionality out of plotnine and have failed so I'm punting.
Someone has to have had good experience with a not-web-oriented locally-hosted plot library that isn't matplotlib. I've seen references to Seaborn, Bokeh, Plotly, Altair, and plotnine. I'd be willing to learn ggplot to leverage the syntax in both R and Python but if there's a decent non-ggplot solution, I'd take it.
There's too goddamn much implicit magic in matplotlib, I have no MATLAB knowledge to leverage, and I'm tired of fighting through the matplotlib learning curve. It's awful and I want something clearer and better.
matplotlib is brittle, unintuitive, and basically unusable. I do not want or need to display plots, I do not want or need web or Jupyter support or any interactivity.
I need something that runs locally and just turns data into lines and points and text in a graphics format that can eventually be printed. Something like what gnuplot does but that can be driven from Python.
Everything either is aimed at "data visualization" on the web or is just a layer on top of matolotlib. I have two columns of numbers, I want a professional-looking B/W scatter or line plot that can be incorporated into a press-friendly document.
plotnine is failing due to some weird-ass matplotlib backend issue that the plotnine maintainer claims is because something is installed wrong. I've spent my alotted hour trying to get the minimum functionality out of plotnine and have failed so I'm punting.
Someone has to have had good experience with a not-web-oriented locally-hosted plot library that isn't matplotlib. I've seen references to Seaborn, Bokeh, Plotly, Altair, and plotnine. I'd be willing to learn ggplot to leverage the syntax in both R and Python but if there's a decent non-ggplot solution, I'd take it.
There's too goddamn much implicit magic in matplotlib, I have no MATLAB knowledge to leverage, and I'm tired of fighting through the matplotlib learning curve. It's awful and I want something clearer and better.