How to make jitterplot on matplolib python

Question

Here is my code (adapted from here):

df_1 = pd.DataFrame({'Cells' : np.arange(0,100), 'Delta_7' : np.random.rand(100,), 'Delta_10' : np.random.rand(100,), 'Delta_14' : np.random.rand(100,)}, columns = ['Cells','Delta_7', 'Delta_10', 'Delta_14'])



#figure
fig, ax1 = plt.subplots()
fig.set_size_inches(13, 10)



#c sequence
c = df_1['Delta_7']

#plot

plt.scatter(np.full((len(df_1), 1), 1), df_1['Delta_7'] , s = 50, c=c, cmap = 'viridis')
plt.scatter(np.full((len(df_1), 1), 2), df_1['Delta_10'] , s = 50, c=c, cmap = 'viridis')
plt.scatter(np.full((len(df_1), 1), 3), df_1['Delta_14'] , s = 50, c=c, cmap = 'viridis')
cbar = plt.colorbar()

I would like to make a beautiful jitterplot (like on R or seaborn) with matplotlib. The thing is that I would like to give each cell a color based on its 'Delta_7' value. And this color would be kept when plotting 'Delta_10' and 'Delta_14', that I didn't manage to do with seaborn. Please, could you let me know if you have any clue (python package, coding tricks …)?

Kindly,

Hi @tomjn! The problem is I do not know how to plot all the 'Delta's column on a same seaborn plot… However, when I tried `ax = sns.stripplot(y="Delta_7", hue='Cells', palette='viridis', data=df_1, dodge=True, jitter=0.2)` all dots have the same color — chalbiophysics, Feb 14 '20 at 16:36

JohanC · Accepted Answer · 2020-02-14T19:53:31.247

The positions of the dots can be obtained from the list returned by scatter. These positions can be jittered, for example only in the x-direction. Possibly the range of the x-axis needs to be extended a bit to show every displaced dot.

Here is some code to start experimenting:

import matplotlib.pyplot as plt
import pandas as pd
import numpy as np

def jitter_dots(dots):
    offsets = dots.get_offsets()
    jittered_offsets = offsets
    # only jitter in the x-direction
    jittered_offsets[:, 0] += np.random.uniform(-0.3, 0.3, offsets.shape[0])
    dots.set_offsets(jittered_offsets)

df_1 = pd.DataFrame({'Cells': np.arange(0, 100),
                     'Delta_7': np.random.rand(100),
                     'Delta_10': np.random.rand(100),
                     'Delta_14': np.random.rand(100)})
fig, ax1 = plt.subplots()

columns = df_1.columns[1:]
c = df_1['Delta_7']
for i, column in enumerate(columns):
    dots = plt.scatter(np.full((len(df_1), 1), i), df_1[column], s=50, c=c, cmap='plasma')
    jitter_dots(dots)
plt.xticks(range(len(columns)), columns)
xmin, xmax = plt.xlim()
plt.xlim(xmin - 0.3, xmax + 0.3)  # make some room to show the jittered dots
cbar = plt.colorbar()
plt.show()

Full pandas noob here... what would I need to change to jitter both in x and y direction? — Sebastian P., May 08 '22 at 00:09
@SebastianP. `jittered_offsets[:, 0] += np.random.uniform(-0.3, 0.3, offsets.shape[0])` changes the x-positions of the dots. `jittered_offsets[:, 1] += np.random.uniform(-0.3, 0.3, offsets.shape[0])` would change the y-positions. — JohanC, May 08 '22 at 22:51

How to make jitterplot on matplolib python

1 Answers1

Linked