2.4. Data Visualisation

Introduction

Effective data visualisation is fundamental to understanding patterns, communicating findings, and making data-driven decisions. This section covers essential visualisation techniques for tabular and spatial data using Python’s powerful plotting libraries.

Before we dive into geospatial visualisation, you need to master the core plotting tools that underpin all visual analysis in Python.

Why Visualisation Matters

Visualisation helps us:

Explore data patterns and relationships
Identify outliers and anomalies
Communicate findings to diverse audiences
Validate analytical results
Guide further analysis and investigation

Anscombe’s Quartet

Four datasets with identical summary statistics but completely different patterns—only visible through visualisation. This classic example demonstrates why we must always plot our data!

Python Visualisation Libraries

The Ecosystem

matplotlib: The foundation - Low-level control over every plot element - Verbose but powerful - Basis for most other plotting libraries

pandas plotting: Quick and convenient - Built-in plotting methods on DataFrames - Good for rapid exploration - Limited customisation

seaborn: Statistical graphics - Beautiful default styles - Statistical functions built-in - High-level interface to matplotlib

plotly: Interactive plots - Web-based interactive graphics - Hover information, zooming, panning - Excellent for dashboards

Installation

# matplotlib and pandas come with most Python installations
import matplotlib.pyplot as plt
import pandas as pd

# Install seaborn if needed
# pip install seaborn
import seaborn as sns

# Install plotly if needed
# pip install plotly
import plotly.express as px

Matplotlib Fundamentals

The Anatomy of a Plot

Every matplotlib plot consists of:

Figure: The entire plotting canvas
Axes: The plotting area with axes, labels, and data
Artists: Everything you see (lines, text, legends)

import matplotlib.pyplot as plt
import numpy as np

# Create figure and axes
fig, ax = plt.subplots(figsize=(8, 6))

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

# Plot
ax.plot(x, y)

# Customize
ax.set_xlabel('X axis')
ax.set_ylabel('Y axis')
ax.set_title('Sine Wave')
ax.grid(True, alpha=0.3)

plt.show()

Basic Plot Types

Line Plot

# Time series or continuous data
fig, ax = plt.subplots(figsize=(10, 6))

dates = pd.date_range('2024-01-01', periods=100)
values = np.cumsum(np.random.randn(100))

ax.plot(dates, values, linewidth=2)
ax.set_xlabel('Date')
ax.set_ylabel('Value')
ax.set_title('Time Series')
plt.xticks(rotation=45)
plt.tight_layout()
plt.show()

Scatter Plot

# Relationship between two variables
fig, ax = plt.subplots(figsize=(8, 6))

x = np.random.randn(100)
y = 2 * x + np.random.randn(100) * 0.5

ax.scatter(x, y, alpha=0.6, s=50)
ax.set_xlabel('X variable')
ax.set_ylabel('Y variable')
ax.set_title('Scatter Plot')
ax.grid(True, alpha=0.3)
plt.show()

Bar Chart

# Categorical data
categories = ['A', 'B', 'C', 'D', 'E']
values = [23, 45, 56, 78, 32]

fig, ax = plt.subplots(figsize=(8, 6))
ax.bar(categories, values, color='steelblue', alpha=0.7)
ax.set_xlabel('Category')
ax.set_ylabel('Value')
ax.set_title('Bar Chart')
plt.show()

Histogram

# Distribution of values
data = np.random.randn(1000)

fig, ax = plt.subplots(figsize=(8, 6))
ax.hist(data, bins=30, edgecolor='black', alpha=0.7)
ax.set_xlabel('Value')
ax.set_ylabel('Frequency')
ax.set_title('Histogram')
plt.show()

Customising Plots

Colours and Styles

# Using named colours
ax.plot(x, y, color='steelblue')
ax.plot(x, y, color='#FF6B6B')  # Hex code

# Line styles
ax.plot(x, y, linestyle='-')   # Solid
ax.plot(x, y, linestyle='--')  # Dashed
ax.plot(x, y, linestyle=':')   # Dotted
ax.plot(x, y, linestyle='-.')  # Dash-dot

# Markers
ax.plot(x, y, marker='o')  # Circle
ax.plot(x, y, marker='s')  # Square
ax.plot(x, y, marker='^')  # Triangle

# Combined
ax.plot(x, y, color='steelblue', linestyle='--', 
        marker='o', markersize=8, linewidth=2)

Multiple Subplots

# Create 2x2 grid of subplots
fig, axes = plt.subplots(2, 2, figsize=(12, 10))

# Access individual axes
axes[0, 0].plot(x, np.sin(x))
axes[0, 0].set_title('Sine')

axes[0, 1].plot(x, np.cos(x))
axes[0, 1].set_title('Cosine')

axes[1, 0].plot(x, np.tan(x))
axes[1, 0].set_title('Tangent')

axes[1, 1].plot(x, x**2)
axes[1, 1].set_title('Quadratic')

plt.tight_layout()
plt.show()

Legends and Annotations

fig, ax = plt.subplots(figsize=(10, 6))

# Multiple lines
ax.plot(x, np.sin(x), label='sin(x)', linewidth=2)
ax.plot(x, np.cos(x), label='cos(x)', linewidth=2)
ax.plot(x, np.tan(x), label='tan(x)', linewidth=2)

# Add legend
ax.legend(loc='upper right', fontsize=12)

# Add text annotation
ax.text(3, 0.5, 'Important point', fontsize=12,
        bbox=dict(boxstyle='round', facecolor='wheat', alpha=0.5))

# Add arrow annotation
ax.annotate('Maximum', xy=(np.pi/2, 1), xytext=(2, 1.5),
            arrowprops=dict(arrowstyle='->', color='red', lw=2),
            fontsize=12)

plt.show()

Pandas Plotting

Built-in Plot Methods

Pandas DataFrames have convenient plotting methods:

# Sample data
df = pd.DataFrame({
    'date': pd.date_range('2024-01-01', periods=100),
    'temperature': 15 + 10 * np.sin(np.linspace(0, 4*np.pi, 100)) + np.random.randn(100),
    'rainfall': np.abs(np.random.randn(100) * 10)
})

# Line plot
df.plot(x='date', y='temperature', figsize=(10, 6))
plt.title('Temperature Over Time')
plt.ylabel('Temperature (°C)')
plt.show()

# Multiple columns
df.set_index('date')[['temperature', 'rainfall']].plot(
    figsize=(12, 6),
    secondary_y='rainfall'  # Rainfall on secondary axis
)
plt.title('Temperature and Rainfall')
plt.show()

# Bar plot
df.groupby(df['date'].dt.month)['rainfall'].sum().plot(kind='bar')
plt.title('Total Rainfall by Month')
plt.xlabel('Month')
plt.ylabel('Total Rainfall (mm)')
plt.show()

# Histogram
df['temperature'].plot(kind='hist', bins=20, edgecolor='black')
plt.title('Temperature Distribution')
plt.xlabel('Temperature (°C)')
plt.show()

# Box plot
df[['temperature', 'rainfall']].plot(kind='box')
plt.title('Distribution Comparison')
plt.show()

Quick Visualisation Methods

# Scatter matrix (pairplot equivalent)
from pandas.plotting import scatter_matrix

scatter_matrix(df[['temperature', 'rainfall']], 
               figsize=(10, 10), 
               alpha=0.5,
               diagonal='kde')
plt.show()

Seaborn: Statistical Visualisation

Why Seaborn?

Beautiful default themes
Statistical functions built-in
Excellent for exploring relationships
Integrates seamlessly with pandas

Setting Styles

import seaborn as sns

# Set style
sns.set_style('whitegrid')  # whitegrid, darkgrid, white, dark, ticks

# Set context
sns.set_context('notebook')  # paper, notebook, talk, poster

# Set colour palette
sns.set_palette('husl')  # or 'Set2', 'colorblind', etc.

Statistical Plots

Distribution Plots

# Histogram with KDE
fig, ax = plt.subplots(figsize=(10, 6))
sns.histplot(data=df, x='temperature', kde=True, ax=ax)
ax.set_title('Temperature Distribution')
plt.show()

# KDE plot
fig, ax = plt.subplots(figsize=(10, 6))
sns.kdeplot(data=df, x='temperature', ax=ax, fill=True)
ax.set_title('Temperature KDE')
plt.show()

# Violin plot
sns.violinplot(data=df, y='temperature')
plt.title('Temperature Violin Plot')
plt.show()

Relationship Plots

# Scatter plot with regression line
sns.lmplot(data=df, x='temperature', y='rainfall', 
           height=6, aspect=1.5)
plt.title('Temperature vs Rainfall')
plt.show()

# Joint plot (scatter + histograms)
sns.jointplot(data=df, x='temperature', y='rainfall',
              kind='scatter', height=8)
plt.show()

# Pair plot
sns.pairplot(df[['temperature', 'rainfall']])
plt.show()

Categorical Plots

# Add categorical variable
df['season'] = pd.cut(df['date'].dt.month, 
                      bins=[0, 3, 6, 9, 12],
                      labels=['Summer', 'Autumn', 'Winter', 'Spring'])

# Box plot by category
fig, ax = plt.subplots(figsize=(10, 6))
sns.boxplot(data=df, x='season', y='temperature', ax=ax)
ax.set_title('Temperature by Season')
plt.show()

# Violin plot by category
fig, ax = plt.subplots(figsize=(10, 6))
sns.violinplot(data=df, x='season', y='temperature', ax=ax)
ax.set_title('Temperature Distribution by Season')
plt.show()

# Swarm plot (individual points)
fig, ax = plt.subplots(figsize=(10, 6))
sns.swarmplot(data=df, x='season', y='temperature', ax=ax, size=3)
ax.set_title('Temperature by Season (All Points)')
plt.show()

Heatmaps

# Correlation matrix
correlation = df[['temperature', 'rainfall']].corr()

fig, ax = plt.subplots(figsize=(8, 6))
sns.heatmap(correlation, annot=True, cmap='coolwarm', 
            center=0, ax=ax, square=True)
ax.set_title('Correlation Matrix')
plt.show()

# Pivot table heatmap
pivot = df.pivot_table(
    values='temperature',
    index=df['date'].dt.month,
    columns=df['date'].dt.day,
    aggfunc='mean'
)

fig, ax = plt.subplots(figsize=(20, 8))
sns.heatmap(pivot, cmap='YlOrRd', ax=ax)
ax.set_title('Temperature: Month vs Day')
ax.set_xlabel('Day of Month')
ax.set_ylabel('Month')
plt.show()

Practical Example: Auckland Weather Analysis

Let’s apply these techniques to real data:

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

# Simulate Auckland weather data
np.random.seed(42)
dates = pd.date_range('2023-01-01', '2023-12-31', freq='D')
n = len(dates)

# Seasonal temperature pattern
day_of_year = dates.dayofyear
temp = 15 + 8 * np.sin(2 * np.pi * (day_of_year - 80) / 365) + np.random.randn(n) * 2

# Rainfall (more in winter)
rainfall_base = 3 - 2 * np.sin(2 * np.pi * (day_of_year - 80) / 365)
rainfall = np.abs(rainfall_base + np.random.randn(n) * 1.5)

# Create DataFrame
weather = pd.DataFrame({
    'date': dates,
    'temperature': temp,
    'rainfall': rainfall,
    'month': dates.month,
    'season': pd.cut(dates.month, 
                     bins=[0, 3, 6, 9, 12],
                     labels=['Summer', 'Autumn', 'Winter', 'Spring'])
})

# 1. Time series overview
fig, axes = plt.subplots(2, 1, figsize=(14, 8), sharex=True)

# Temperature
axes[0].plot(weather['date'], weather['temperature'], 
             linewidth=1, alpha=0.7, color='orangered')
axes[0].set_ylabel('Temperature (°C)', fontsize=12)
axes[0].set_title('Auckland Weather 2023', fontsize=14, fontweight='bold')
axes[0].grid(True, alpha=0.3)

# Rainfall
axes[1].bar(weather['date'], weather['rainfall'], 
            width=1, color='steelblue', alpha=0.6)
axes[1].set_ylabel('Rainfall (mm)', fontsize=12)
axes[1].set_xlabel('Date', fontsize=12)
axes[1].grid(True, alpha=0.3)

plt.tight_layout()
plt.show()

# 2. Seasonal comparison
fig, axes = plt.subplots(1, 2, figsize=(14, 6))

sns.boxplot(data=weather, x='season', y='temperature', ax=axes[0])
axes[0].set_title('Temperature by Season', fontsize=12)
axes[0].set_ylabel('Temperature (°C)')

sns.boxplot(data=weather, x='season', y='rainfall', ax=axes[1])
axes[1].set_title('Rainfall by Season', fontsize=12)
axes[1].set_ylabel('Rainfall (mm)')

plt.tight_layout()
plt.show()

# 3. Distributions
fig, axes = plt.subplots(1, 2, figsize=(14, 6))

sns.histplot(data=weather, x='temperature', kde=True, ax=axes[0])
axes[0].set_title('Temperature Distribution', fontsize=12)
axes[0].axvline(weather['temperature'].mean(), 
                color='red', linestyle='--', label='Mean')
axes[0].legend()

sns.histplot(data=weather, x='rainfall', kde=True, ax=axes[1])
axes[1].set_title('Rainfall Distribution', fontsize=12)
axes[1].axvline(weather['rainfall'].mean(), 
                color='red', linestyle='--', label='Mean')
axes[1].legend()

plt.tight_layout()
plt.show()

# 4. Relationship
fig, ax = plt.subplots(figsize=(10, 6))
sns.scatterplot(data=weather, x='temperature', y='rainfall', 
                hue='season', alpha=0.6, ax=ax)
ax.set_title('Temperature vs Rainfall by Season', fontsize=12)
ax.set_xlabel('Temperature (°C)')
ax.set_ylabel('Rainfall (mm)')
plt.show()

# 5. Monthly summary
monthly = weather.groupby('month').agg({
    'temperature': ['mean', 'std'],
    'rainfall': 'sum'
}).round(2)

print("\nMonthly Summary:")
print(monthly)

Best Practices

Design Principles

1. Choose the Right Plot Type - Line plots: Trends over time - Bar charts: Comparing categories - Scatter plots: Relationships between variables - Histograms: Distribution of single variable - Box plots: Distribution comparisons

2. Keep It Simple - One main message per plot - Remove unnecessary elements - Use clear labels and titles - Choose appropriate colour scales

3. Make It Readable - Sufficient font sizes (≥10pt) - Adequate spacing - Clear axis labels with units - Informative titles

4. Use Colour Effectively - Colourblind-friendly palettes - Consistent colour meanings - Appropriate for data type (sequential, diverging, categorical)

Common Pitfalls to Avoid

❌ Don’t: - Use 3D charts (they distort perception) - Start y-axis at non-zero without good reason - Use too many colours - Overload with information - Use defaults without thought

✓ Do: - Show data clearly - Label everything - Use appropriate scales - Consider your audience - Test with others

Saving Figures

# High-resolution for publications
fig.savefig('plot.png', dpi=300, bbox_inches='tight')

# Vector format for editing
fig.savefig('plot.pdf', bbox_inches='tight')
fig.savefig('plot.svg', bbox_inches='tight')

# Transparent background
fig.savefig('plot.png', dpi=300, bbox_inches='tight', transparent=True)

Summary

In this section, you’ve learned:

Matplotlib fundamentals: Figure, axes, and basic plot types
Pandas plotting: Quick visualisations from DataFrames
Seaborn: Beautiful statistical graphics
Customisation: Colours, styles, legends, annotations
Best practices: Choosing plot types and designing effective visualisations

These skills form the foundation for all data visualisation in Python, including the geospatial visualisation you’ll learn in the next section (sec-geospatial).

Practice Exercises

Explore a Dataset: Load a CSV file and create 5 different visualisations showing different aspects of the data
Time Series Analysis: Create a multi-panel plot showing temperature, rainfall, and a derived metric (e.g., comfort index)
Correlation Analysis: Create a correlation heatmap for a multi-variable dataset and identify strong relationships
Distribution Comparison: Compare distributions across categories using box plots, violin plots, and histograms
Custom Dashboard: Create a 2x2 subplot figure summarising a dataset with different visualisation types

Next Steps

Now that you understand general data visualisation, you’re ready to tackle geospatial visualisation in sec-geospatial, where you’ll learn to create maps and spatial graphics.

Introduction

Why Visualisation Matters

Python Visualisation Libraries

The Ecosystem

Installation

Matplotlib Fundamentals

The Anatomy of a Plot

Basic Plot Types

Customising Plots

Colours and Styles

Multiple Subplots

Legends and Annotations

Pandas Plotting

Built-in Plot Methods

Quick Visualisation Methods

Seaborn: Statistical Visualisation

Why Seaborn?

Setting Styles

Statistical Plots

Practical Example: Auckland Weather Analysis

Best Practices

Design Principles

Common Pitfalls to Avoid

Saving Figures

Summary

Practice Exercises

Further Reading

Next Steps