Dimensione esadecimale in hexbins matplotlib in base alla densità dei punti vicini

Ho il seguente codice che produce la figura seguente

import numpy as np
np.random.seed(3)
import pandas as pd
import matplotlib.pyplot as plt
df = pd.DataFrame()
df['X'] = list(np.random.randint(100, size=100)) + list(np.random.randint(30, size=100))
df['Y'] = list(np.random.randint(100, size=100)) + list(np.random.randint(30, size=100))

df['Bin'] = df.apply(lambda row: .1 if row['X'] < 30 and row['Y'] < 30 else .9, axis=1)

fig, ax = plt.subplots(figsize=(10,10))
plt.scatter(df['X'], df['Y'])

Ho rappresentato graficamente i dati usando hexbin, come indicato di seguito

from matplotlib import cm

fig, ax = plt.subplots(figsize=(10,10))
hexbin = ax.hexbin(df['X'], df['Y'], C=df['Bin'], gridsize=20, cmap= cm.get_cmap('RdYlBu_r'),edgecolors='black')
plt.show()

Vorrei cambiare la dimensione degli esagoni in base alla densità dei punti tracciati nell'area coperta da un esagono. Ad esempio, gli esagoni in basso a sinistra (dove i punti sono compatti) saranno più grandi degli esagoni ovunque (dove i punti sono radi). C'è un modo per fare questo?

Modifica: ho provato questa soluzione , ma non riesco a capire come colorare gli esagoni in base a df ['Bin'] o come impostare la dimensione esadecimale minima e massima.

from matplotlib.collections import PatchCollection
from matplotlib.path import Path
from matplotlib.patches import PathPatch
fig, ax = plt.subplots(figsize=(10,10))
hexbin = ax.hexbin(df['X'], df['Y'], C=df['Bins'], gridsize=20, cmap= cm.get_cmap('RdYlBu_r'),edgecolors='black')
def sized_hexbin(ax,hc):
    offsets = hc.get_offsets()
    orgpath = hc.get_paths()[0]
    verts = orgpath.vertices
    values = hc.get_array()
    ma = values.max()
    patches = []
    for offset,val in zip(offsets,values):
        v1 = verts*val/ma+offset
        path = Path(v1, orgpath.codes)
        patch = PathPatch(path)
        patches.append(patch)

    pc = PatchCollection(patches, cmap=cm.get_cmap('RdYlBu_r'), edgecolors='black')
    pc.set_array(values)
    ax.add_collection(pc)
    hc.remove()

sized_hexbin(ax,hexbin)
plt.show()

Potresti voler trascorrere qualche tempo a comprendere la mappatura dei colori.

    import numpy as np
    np.random.seed(3)
    import pandas as pd
    import matplotlib.pyplot as plt
    from matplotlib.collections import PatchCollection
    from matplotlib.path import Path
    from matplotlib.patches import PathPatch
    df = pd.DataFrame()
    df['X'] = list(np.random.randint(100, size=100)) + list(np.random.randint(30, size=100))
    df['Y'] = list(np.random.randint(100, size=100)) + list(np.random.randint(30, size=100))

    df['Bin'] = df.apply(lambda row: .1 if row['X'] < 30 and row['Y'] < 30 else .9, axis=1)

    #fig, ((ax1, ax2)) = plt.subplots(1, 2, sharex=True, sharey=True)
    ax1 = plt.scatter(df['X'], df['Y'])

    fig,ax2 = plt.subplots(figsize=(10,10))
    hexbin = ax2.hexbin(df['X'], df['Y'], C=df['Bin'], gridsize=20,edgecolors='black',cmap= 'RdBu', reduce_C_function=np.bincount) #**

    def sized_hexbin(ax,hc):
        offsets = hc.get_offsets()
        orgpath = hc.get_paths()[0]
        verts = orgpath.vertices
        values = hc.get_array()
        ma = values.max()
        patches = []
        for offset,val in zip(offsets,values):
            v1 = verts*val/ma + offset
            path = Path(v1, orgpath.codes)
            patch = PathPatch(path)
            patches.append(patch)

        pc = PatchCollection(patches, cmap= 'RdBu', edgecolors='black')
        pc.set_array(values)

        ax.add_collection(pc)

        hc.remove()

    sized_hexbin(ax2,hexbin)
    cb = plt.colorbar(hexbin, ax=ax2)

    plt.show()

To plot the chart based on df['bins'] values - 

Need to change the reduce_C_function in #** marked line -

    hexbin = ax2.hexbin(df['X'], df['Y'], C=df['Bin'], gridsize=20,edgecolors='black',cmap= 'RdBu', reduce_C_function=np.sum)

[![enter image description here][2]][2]


  [1]: https://i.stack.imgur.com/kv0U4.png
  [2]: https://i.stack.imgur.com/mb0gD.png

# Another variation of the chart :

# Where size is based on count of points in the bins and color is based on values of the df['bin']./ Also added if condition to control minimum hexbin size.


import numpy as np
np.random.seed(3)
import pandas as pd
import matplotlib.pyplot as plt
from matplotlib.collections import PatchCollection
from matplotlib.path import Path
from matplotlib.patches import PathPatch
from functools import partial

mycmp = 'coolwarm'

df = pd.DataFrame()
df['X'] = list(np.random.randint(100, size=100)) + list(np.random.randint(30, size=100))
df['Y'] = list(np.random.randint(100, size=100)) + list(np.random.randint(30, size=100))

df['Bin'] = df.apply(lambda row: .1 if row['X'] < 30 and row['Y'] < 30 else .9, axis=1)

#fig, ((ax1, ax2)) = plt.subplots(1, 2, sharex=True, sharey=True)
ax1 = plt.scatter(df['X'], df['Y'])


fig,ax2 = plt.subplots(figsize=(10,10))
hexbin = ax2.hexbin(df['X'], df['Y'], C=df['Bin'], gridsize=15,edgecolors='black',cmap= newcmp , reduce_C_function=np.bincount)
hexbin2 = ax2.hexbin(df['X'], df['Y'], C=df['Bin'], gridsize=15,edgecolors='black',cmap= newcmp , reduce_C_function=np.mean)

def sized_hexbin(ax,hc,hc2):
    offsets = hc.get_offsets()
    orgpath = hc.get_paths()[0]
    verts = orgpath.vertices
    values1 = hc.get_array()
    values2 = hc2.get_array()
    ma = values1.max()
    patches = []

    for offset,val in zip(offsets,values1):
        # Adding condition for minimum size 
        if (val/ma) < 0.2:
            val_t = 0.2
        else:
            val_t = val/ma
        v1 =  verts*val_t + offset
        path = Path(v1, orgpath.codes)
        print(path)
        patch = PathPatch(path)
        patches.append(patch)

    pc = PatchCollection(patches, cmap=  newcmp)  #edgecolors='black'
    pc.set_array(values2)

    ax.add_collection(pc)
    hc.remove()
    hc2.remove()


sized_hexbin(ax2,hexbin,hexbin2)
cb = plt.colorbar(hexbin2, ax=ax2)

plt.xlim((-5, 100))
plt.ylim((-5, 100))

plt.show()