如何将特定轴添加到matplotlib子图?

时间:2016-07-25 15:32:33

标签: python matplotlib subplot

我正在尝试使用matplotlib制作矩阵图。

单个图是使用特定模块windrose制作的,其中PolarAxes是子类。但是,模块中似乎没有定义任何投影作为子图kwargs。标准polar投影不起作用,因为缺少某些子类参数。

我已经测试了几种方法但没有成功(即使考虑到这篇文章的seaborn地图:https://stackoverflow.com/a/25702476/3416205)。以下是我最接近的尝试。如果没有正确创建与特定WindroseAxes相关联的新matplotlib投影,有没有办法做我想做的事情?

import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.gridspec as gridspec
from windrose import WindroseAxes

df = pd.read_csv('https://raw.githubusercontent.com/AntoineGautier/Data/master/tmp.csv')
fig = plt.figure()
gs = gridspec.GridSpec(4, 2)

def wind_plot(x, y, title=None, axes=None, fig=None):
    ax = WindroseAxes.from_ax()
    ax.set_position(axes.get_position(fig))
    ax.bar(x, y, normed=True, opening=0.8, edgecolor='white', bins=[0, 2.5, 5, 7.5, 10])
    ax.set_title(title)

for (id_s, s) in enumerate(pd.unique(df.saison)):
    for (id_jn, jn) in enumerate(pd.unique(df.jn)):
        tmp = df.query('saison==@s & jn==@jn')
        _ = plt.subplot(gs[id_s, id_jn], polar=True)
        wind_plot(tmp.wd, tmp.ws, title=s + ' - ' + jn, axes=_, fig=fig)

plt.show()

1 个答案:

答案 0 :(得分:1)

windrose模块中的github页面实际上提供了子图的示例:https://github.com/scls19fr/windrose/blob/master/samples/example_subplots.py

以下作品。

import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.gridspec as gridspec
from windrose import WindroseAxes

df = pd.read_csv('https://raw.githubusercontent.com/AntoineGautier/Data/master/tmp.csv')

fig = plt.figure(figsize=(20, 10))
gs = gridspec.GridSpec(2, 4)
gp = gs.get_grid_positions(fig)  # [bottom, top, left, right]

def wind_plot(x, y, title, fig, rect):
    ax = WindroseAxes(fig, rect)
    fig.add_axes(ax)
    ax.bar(x, y, normed=True, opening=0.8, edgecolor='white', bins=[0, 2.5, 5, 7.5, 10])
    ax.set_title(title, position=(0.5, 1.1))

for (id_s, s) in enumerate(pd.unique(df.saison)):
    for (id_jn, jn) in enumerate(pd.unique(df.jn)):
        tmp = df.query('saison==@s & jn==@jn')
        rect = [gp[2][id_s], gp[0][id_jn],
                gp[3][id_s]-gp[2][id_s],
                gp[1][id_jn]-gp[0][id_jn]]  # [left, bottom, width, height]
        wind_plot(tmp.wd, tmp.ws, s + ' | ' + jn, fig, rect)

plt.show()