CODEWITHSUNDEEP
pythonpandasplt
Anjo
Anjo

Reputation: 175

Scatter plots in Pandas: Plot by category with different color and shape combinations

I would like to plot a data-set by its categories, using geometric shapes such as circle, triangle and square to represent category 1 and colors to represent category 2. The output would have varying combination of the geometric shapes and colors and the legend would list the attributes of the categories separately i.e.:

circle = a
triangle = b
square = c

red = I
green = II
blue = III

Looking for solutions I found following posts which would only give solutions for one specific geometric shape having one specific color.

I tried to work something out with the code from one of the posts but without success.

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

np.random.seed(1983)
num = 10
x, y = np.random.random((2, num))
cat1 = np.random.choice(['a', 'b', 'c'], num)
cat2 = np.random.choice(['I', 'II', 'III'], num)
df = pd.DataFrame(dict(x=x, y=y, cat1=cat1, cat2=cat2))

groups = df.groupby(['cat1', 'cat2'])

fig, ax = plt.subplots()
for name, group in groups:
ax.plot(group.x, group.y, marker='o', linestyle='', ms=12, label=name)
ax.legend()

plt.show()

Upvotes: 1

Views: 3122

Answers (1)

Aakash Makwana
Aakash Makwana

Reputation: 754

you can try this code block

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

#Create mapping dictionary that you want
marker_dict = {'a':'o','b':'^','c':'s'}
color_dict = {'I':'red', 'II':'green', 'III':'blue'}

np.random.seed(1983)
num = 10
x, y = np.random.random((2, num))
cat1 = np.random.choice(['a', 'b', 'c'], num)
cat2 = np.random.choice(['I', 'II', 'III'], num)
df = pd.DataFrame(dict(x=x, y=y, cat1=cat1, cat2=cat2))

groups = df.groupby(['cat1', 'cat2'])

fig, ax = plt.subplots()
ax.margins(0.05)
for name, group in groups:
    marker = marker_dict[name[0]]
    color = color_dict[name[1]]
    ax.plot(group.x, group.y, marker=marker, linestyle='', ms=12, label=name,color=color)
ax.legend()

plt.show()

Hope it helps.

Upvotes: 3

Related Questions