Spread the love

Learn how to make beautiful GUI apps

Comprehensive, for Kivy beginners, easy to follow.

Get the book here (PDF) or on Amazon:

ebook / paperback (black and white) / paperback (full color)

Scatter Plots with the scatter Method

In the previous article we were using the plot method to create scatter plots. But there’s a specialized and more powerful method to do that. In this article we’re going to be using it. Let’s have a look at the example from the previous article:

In [1]:

%matplotlib inline
import matplotlib as mpl
import matplotlib.pyplot as plt
import numpy as np

x = np.linspace(0, 10, 40)
y = np.sin(x)

fig, ax = plt.subplots()
ax.plot(x, y, 'o')

Out[1]:

[<matplotlib.lines.Line2D at 0x1c4ac899688>]

Now let’s use the scatter method instead of the plot method. Here we must explicitly specify what the third argument is:

In [2]:

fig, ax = plt.subplots()
ax.scatter(x, y, marker='o')

Out[2]:

<matplotlib.collections.PathCollection at 0x1c4ae112188>

You can also use other markers:

In [3]:

fig, ax = plt.subplots()
ax.scatter(x, y, marker='s')

Out[3]:

<matplotlib.collections.PathCollection at 0x1c4ae116108>

Unlike the plot method, scatter lets us control each point individually.

In [4]:

# Here's how we can control the sizes of the points. Let's define a list 
# of sizes from smallest to greatest using a list comprehension.
sizes = [i for i in range(40)]

# And now let's use the sizes to plot the function.
fig, ax = plt.subplots()
ax.scatter(x, y, s=sizes)

Out[4]:

<matplotlib.collections.PathCollection at 0x1c4aed82688>

Here’s another example:

In [6]:

# Let's randomize the sizes. This code will produce a different output each time you execute it.
sizes = [i for i in 100 * np.random.randn(10)]
fig, ax = plt.subplots()
ax.scatter(x, y, s=sizes, marker='<')

Out[6]:

<matplotlib.collections.PathCollection at 0x1c4aeeabdc8>

We can also control the colors of the markers:

In [7]:

# Let's create a list of colors and use it in the scatter method.
# Here we're using 4 colors. As there are 40 points, we multiply  
# the list by 10 to repeat its elements 10 times, thus getting 
# 40 elements altogether.
colors = 10 * ['b', 'r', 'g', 'y']
fig, ax = plt.subplots()
ax.scatter(x, y, c=colors)

Out[7]:

<matplotlib.collections.PathCollection at 0x1c4aeee9bc8>

Here’s an example with random colors:

In [8]:

# This code will produce a different output each time you execute it.
colors = np.random.randn(40)
fig, ax = plt.subplots()
ax.scatter(x, y, c=colors)

Out[8]:

<matplotlib.collections.PathCollection at 0x1c4aef5ff08>

You can also set the alpha value:

In [9]:

# Here the points are pretty big, so you can watch them overlap.
fig, ax = plt.subplots()
ax.scatter(x, y, marker='o', s=1000, alpha=.5)

Out[9]:

<matplotlib.collections.PathCollection at 0x1c4aefce308>

If you can achieve your goal with the plot method, you should choose it over scatter because it’s more efficient. It’s especially important for large datasets.