Data Science Scatter Plot

What a Scatter Plot Shows

Every point on a scatter plot is one observation, positioned by its x-value on the horizontal axis and its y-value on the vertical axis. With enough points, the overall shape of the cloud reveals whether the two variables are related.

Creating a Scatter Plot

matplotlib.pyplot.scatter takes two equal-length arrays - one for the x-coordinates, one for the y-coordinates - and draws one dot per pair.

import numpy as np
import matplotlib.pyplot as plt

# hours of exercise per week vs resting heart rate (beats per minute)
exercise_hours = np.array([0, 1, 2, 3, 4, 5, 6, 7, 8])
resting_hr = np.array([82, 80, 78, 74, 73, 70, 68, 65, 64])

plt.scatter(exercise_hours, resting_hr)
plt.xlabel("Exercise hours per week")
plt.ylabel("Resting heart rate (bpm)")
plt.show()

The points here trend steadily downward: more weekly exercise lines up with a lower resting heart rate. That downward drift is what a negative correlation looks like on a scatter plot.

Reading Correlation From the Shape

• A trend rising left to right (positive correlation)
• A trend falling left to right (negative correlation)
• A cloud with no visible trend (little or no correlation)
• A bend or curve rather than a straight trend (a nonlinear relationship)

import numpy as np
import matplotlib.pyplot as plt

np.random.seed(0)
x = np.random.rand(50) * 10
y = np.random.rand(50) * 10

plt.scatter(x, y)
plt.xlabel("x")
plt.ylabel("y")
plt.show()

Once a scatter plot shows a roughly straight-line trend, you can go a step further and fit an actual line through the points - that's exactly what linear regression, covered next, does.