将每个数组的散点移动 delta x

I'm trying to sort out a plot which at the moment looks like this:

我正在尝试找出如何针对不同的情况进行轮班dx每个数据集的值，在这种模式下，系列的最后一项保持在中心（在这种情况下q），而前 8 个项目（从 a 到 h）将向左移动，后 8 个向右移动（从 g 到 p），每个项目都有不同的偏移量。

我有两个形状为 (12, 17) 的数组（M 和 M1）对于循环中的每个项目，因此该形状对应于one散点的颜色。

    import matplotlib.cm as cm
    import matplotlib
    import matplotlib.pyplot as plt
    import numpy as np

    fig, ax0 = plt.subplots (nrows=1, ncols=1)

    months_expanded = np.zeros((12,17))  
    months = np.arange(1, 13)
    M  = np.random.rand(12,17)
    M1 = np.random.rand(12,17)
    datalist = ['a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q']

    for idx, sub in enumerate(datalist):
        for i in range(0,12):
            months_expanded[i] = np.repeat(months[i], 17)

        difference = abs(M - M1)

        ax0.scatter(months_expanded[:, idx], difference[:, idx], label = sub)                       

        colormap = plt.cm.gist_ncar                                                 
        colorst = [colormap(i) for i in    np.linspace(0,0.9,len(ax0.collections))]       
        for t,j1 in enumerate(ax0.collections):
            j1.set_color(colorst[t])

       ax0.legend(loc='center left', bbox_to_anchor=(1, 0.5))
       ax0.yaxis.grid(True)
       ax0.xaxis.grid(True)

       matplotlib.rcParams.update({'font.size': 30})
    fig.set_size_inches(20,20)
    fig.savefig(outfile, bbox_inches='tight')

您可以预定义偏移并创建一组包含偏移的 x 值。请参阅下面的最小示例来说明这一点：

import numpy as np
import matplotlib.pyplot as plt

N = 5 # This would be 17 in your case

# Create offsets
dx = np.linspace(-0.5, 0.5, N)
# Put last data set in centre
dx[N//2:] += (dx[1]-dx[0])
dx[-1] = 0.0
dx_expanded = np.repeat(dx, 12).reshape(N, 12)

# X-values to be shifted
months = np.arange(1, 13)
months_expanded = np.tile(months, N).reshape(N, 12)
months_shifted = months_expanded + dx_expanded

# Y-values
M  = np.random.rand(12,N)
M = np.linspace(1, 5, 12)
M = np.tile(M, N).reshape(N,12)

# Plot results
for idx in range(N):
  plt.scatter(months_shifted[idx], M[idx])
plt.show()

Result: