Plots are typical composed, and when writing a paper (I insert them mostly into TeX publications) I do find the quality of the resulting plot is just so much more refined.
Seaborn is indeed closer and was definitely inspired by ggplot2 in some areas, but IMHO, it’s still not 100% there visually. I’m very much a Python user and would love it to be, but when I’m, let’s say, publishing a book, I’d always go back to ggplot2 - when preparing a paper for a lab class, seaborn is probably fine.
It’s a lot more like Seaborn. It produces gorgeous plots with a lovely syntax that is quick and easy to use, but it’s not a full drawing toolkit like matplotlib.
If I need the plot to have a very precise aesthetic, mpl is great. But if I want a high quality statistical plot that looks great. ggplot2 will do it in about 2 seconds. See also plotnine.
I have no idea how op thinks they could make a decent histogram any quicker than ggplot(data) + geom_histogram(x= x). I mean you don’t even have to leave your shell/editor or extract the SQL into CSV.
How does it compare to matplotlib?
Plots are typical composed, and when writing a paper (I insert them mostly into TeX publications) I do find the quality of the resulting plot is just so much more refined.
Seaborn is indeed closer and was definitely inspired by ggplot2 in some areas, but IMHO, it’s still not 100% there visually. I’m very much a Python user and would love it to be, but when I’m, let’s say, publishing a book, I’d always go back to ggplot2 - when preparing a paper for a lab class, seaborn is probably fine.
Same here. I mostly work with Python but the graphs? They are ggplot2.
Plotnine is getting there
It’s a lot more like Seaborn. It produces gorgeous plots with a lovely syntax that is quick and easy to use, but it’s not a full drawing toolkit like matplotlib.
If I need the plot to have a very precise aesthetic, mpl is great. But if I want a high quality statistical plot that looks great. ggplot2 will do it in about 2 seconds. See also plotnine.
I have no idea how op thinks they could make a decent histogram any quicker than
ggplot(data) + geom_histogram(x= x). I mean you don’t even have to leave your shell/editor or extract the SQL into CSV.