Day 31 of 50 Days of Python: Data Visualisation with Matplotlib

Part of Week 5: Data Analysis and Visualisation

Hello again, and welcome to Day 31! We’ll be covering matplotlib and how to generate different visualisations as well as how to customise them with different colours. This is a good lesson as it helps you understand the ins and outs of what matplotlib is capable of.

---

Python Setup

We did cover this in our last lesson but just incase, you can install matplotlib with this line of code:

pip install matplotlib

---

Choosing the Right Visualisation

Different datasets require different visualisation techniques. Here’s a quick guide for choosing the right one:

• Time Series Data

  • Plot types: Line and area

  • Imagine you want to visualise stock prices over time, a line chart can perfectly show this.

• Categorical Comparison

  • Plot types: Bar and Horizontal Bar

  • A great way to show sales by product category or subcategory

• Distribution

  • Plot types: Histogram, box, violin

  • Lets say you want to visualise age distrubtion of a population by country, county or city. These plots are great for showing this.

• Relationship Between Variables

  • Plot types: Scatter, bubble, Pair

  • Top choices for visualising data like the correlation between spend and sales.

• Proportional Data

  • Plot types: Scatter and Stacked Bar

  • Market share distribution is a good example of when to use these plots.

• Geospatial Data

  • Plot types: Heatmap

  • Great for showing population density across regions. I’m sure we’ve all seen a fair few of these especially back in school.

---

---

Customising Your Plots

Matplotlib offers extensive customization options to enhance the clarity and appeal of your visualizations.

Colour Customisation

Specify colours using various formats:

• Named Colours: 'red', 'blue', etc.

• Hexadecimal: '#FF5733'

• RGB Tuple: (0.1, 0.2, 0.5)

• Colormap: 'viridis', 'plasma', etc.

import matplotlib.pyplot as plt

# Line plot with customised colour
plt.plot([1, 2, 3, 4], [10, 20, 25, 30], color='skyblue')
plt.show()

If you’re after a list of colours to use, you can get this from matplotlib documentation here.

Colour Maps

Matplotlib also has a list of a few colour maps for different types of data.

• Sequential: For ordered data (e.g., 'Blues', 'Purples')

• Diverging: For data with a critical midpoint (e.g., 'RdBu', 'PiYG')

• Qualitative: For categorical data (e.g., 'Set1', 'Pastel1')

import numpy as np

# Create data
x = np.linspace(0, 10, 100)
y = np.sin(x)

# Scatter plot with colormap
plt.scatter(x, y, c=y, cmap='viridis')
plt.colorbar()  # Show color scale
plt.show()

Plot Styles

# Set style
plt.style.use('ggplot')

# Create plot
plt.plot([1, 2, 3, 4], [10, 20, 25, 30])
plt.show()

Again, depending on exactly what you’re after you can see what plot styles matplotlib has via their documentation here.

---

Next Up: Day 32 - Advanced Visualisation with Seaborn.

Since we’ve covered matplotlib it would be rude not to at least look at what seaborn is capable of. So Day 32 will follow the same format. we’ll cover what seaborn can do and maybe you’ll use that in place of matplotlib or even as well as.

Hope today has been informative, and as always… Happy coding!