Using numpy with Matplotlib
Martin McBride, 2022-06-10
Tags numeric python numpy linspace function plot pandas matplotlib scipy data science
Categories matplotlib numpy

The previous article showed how to create a plot using Python lists to store the data. Matplotlib also works with NumPy, and this can often simplify things.
Here we will look at how to use numpy to create the x values and y values more easily.
Creating the x values
We want to create a set of 100 x values, equally spaced over the range 0 to 12. We can do this using Python lists by scaling the loop variable.
However, numpy has a function linspace that is designed to do this job. It has two advantages:
- it is more readable
- it covers the exact range 0 to 12, with the intermediate values equally spaced
If you recall, our previous code went from 0 to almost 12, which wasn't really much of a problem, but it wasn't quite correct.
Here is how we use linspace
to create our range:
xa = np.linspace(0, 12, 100)
Creating the y values
NumPy supports universal functions. These allow you to operate on entire arrays in one go:
ya = np.sin(xa)*np.exp(-xa/4)
This performs the same calculation on each of the 100 elements of xa
, creating a new 100-element array ya
containing the results. This avoids a loop, which makes the code more readable, but also more efficient.
Complete code
Here is the complete code:
from matplotlib import pyplot as plt import numpy as np xa = np.linspace(0, 12, 100) ya = np.sin(xa)*np.exp(-xa/4) plt.plot(xa, ya) plt.show()
This code is shorter and more readable, as well as being more efficient. This becomes more important with large data sets, and especially 2-dimensional data.
Here is the output:
This image is almost exactly the same as before. There is a very tiny change in scale because (as we noted before) this graph has values 0 to 12 whereas the previous graph has values 0 to almost 12. The difference is only visible if you compare the graphs pixel for pixel.
Matplotlib for data science
NumPy is a key component used in data science applications, along with Pandas and SciPy. Matplotlib works well with these libraries, making it a very useful library for data science visualisation.
The code for this section is available on github as numpy-function.py.