Post AT9NHdQW4v4v62oKhs by gekitsu@toot.cat
 (DIR) More posts by gekitsu@toot.cat
 (DIR) Post #AT9K2ApnnkN4DHZeQy by john@sauropods.win
       2023-02-28T20:30:17Z
       
       0 likes, 0 repeats
       
       So I'm doing some graphs with Python, and I’m trying to use Matplotlib. I feel like it's pretty impossible to use. Am I crazy? Is there something better out there I could be using?#programming #python
       
 (DIR) Post #AT9KME7rH4NoDEa7eK by basepair@vis.social
       2023-02-28T20:33:57Z
       
       0 likes, 0 repeats
       
       @johnThere are e.g. seaborn and plotnine. They differ quite a lot in terms of how you compose the graph. Both can achieve quite a lot but I suppose it's a matter of taste what works best for you.
       
 (DIR) Post #AT9KkwD91AtntGpyaG by john@sauropods.win
       2023-02-28T20:38:25Z
       
       0 likes, 0 repeats
       
       @basepair I'm actually after something simple. A single data series, dates against a number. I feel like I'm using rocket ships to get to the local shop.
       
 (DIR) Post #AT9L8iui6nZZbEYa8G by basepair@vis.social
       2023-02-28T20:42:31Z
       
       0 likes, 0 repeats
       
       @johnIf programming is not a requirement, then https://www.rawgraphs.io/ is one of the best websites to paste some data and get great graphs with a ton of options/customisation. Bonus: it's open source
       
 (DIR) Post #AT9NHdQW4v4v62oKhs by gekitsu@toot.cat
       2023-02-28T21:06:47Z
       
       0 likes, 0 repeats
       
       @john yeah, matplotlib is clunky as heck and very involved to use. on the up side, if you put your mind to it, you can plot pretty much whatever you wish. (i once used it for comparisons in bending profile between different bow limb designs. it caused me a lot of pain and i’m still discovering new grey hairs i must’ve grown that day.)my everyday go-to is seaborn, which produces pretty graphs with more or less a single call, as long as what you have matches the way it wants its data.
       
 (DIR) Post #AT9Ni5sCw9giqqNbYe by stefan_hessbrueggen@fedihum.org
       2023-02-28T21:11:30Z
       
       0 likes, 0 repeats
       
       @john yes and no: https://www.anaconda.com/blog/python-data-visualization-2018-why-so-many-libraries
       
 (DIR) Post #AT9O0oYVY2mAdx5lXU by neia@are.waffles.fun
       2023-02-28T21:14:57Z
       
       0 likes, 0 repeats
       
       @john Can you call gnuplot from the command line instead?
       
 (DIR) Post #AT9OJWn9ojkJ3IrqQC by pmcarlton@fediscience.org
       2023-02-28T21:18:13Z
       
       0 likes, 0 repeats
       
       @john @futurebird I find that plotly.express is great for 95% of what I want to do. Just pip install dash, import plotly.express as px, then eg. px.line(x=year, y=value). That will open a browser window on your machine with a fully interactive plot.
       
 (DIR) Post #AT9Od0CkgYOhmD7pOC by davidr@hachyderm.io
       2023-02-28T21:21:45Z
       
       0 likes, 0 repeats
       
       @john You aren't crazy. #maplotlib is incredibly complicated and it feels like 75% of that complication is due to trying to be simple. "We'll just make two completely different interfaces but not make that clear to anyone and when people answer questions online you'll get a mix! Perfect!"Most simpler alternatives run matplotlib under the hood and, despite what most people say, this just increases complexity in the long run.
       
 (DIR) Post #AT9QaD1wk82GauroOG by tncowart@assemblag.es
       2023-02-28T21:43:43Z
       
       0 likes, 0 repeats
       
       @john you are not crazy, matplotlib has truly awful apis.
       
 (DIR) Post #AT9u6qFnLPEUBHG05Q by janeadams@vis.social
       2023-03-01T03:14:39Z
       
       0 likes, 0 repeats
       
       @john Echoing what others have suggested, Plotly expressions great for the basics. Altair also has a very nice syntax 👌