Chapter 4: Data Visualization 📈

You've gathered your data with SQL, cleaned and manipulated it with Python and Pandas. Now what? Raw numbers and tables can be hard to understand. **Data Visualization** is the art and science of translating data into visual context, like charts and graphs, to make it easier for the human brain to grasp patterns, trends, and outliers.

For a data analyst, visualization is a critical skill for both **exploratory analysis** (understanding the data yourself) and **communication** (presenting findings to others). A good chart can tell a story much more effectively than a spreadsheet.

Theory: Why Visualize Data? 🤔

• **Pattern Recognition:** Our brains are wired to process visual information quickly. Charts make it easier to spot trends, relationships, and unusual data points that might be hidden in tables.
• **Communication:** Visualizations provide a clear, concise way to communicate complex findings to both technical and non-technical audiences.
• **Storytelling:** Well-designed visuals can guide your audience through the data and highlight the key insights, forming a compelling narrative.
• **Exploration:** Creating quick plots during analysis helps you understand the data's distribution, identify potential issues (like outliers), and guide your next steps.

Anscombe's Quartet is a famous example: four datasets with nearly identical simple descriptive statistics, yet they look vastly different when plotted, highlighting the importance of visualization beyond just summary numbers.

Theory: Principles of Effective Visualization ✨

Creating a chart is easy; creating an *effective* chart requires thought:

• **Choose the Right Chart Type:** Different charts are suited for different types of data and questions (more on this below).
• **Clarity and Simplicity:** Avoid clutter ("chart junk"). The visualization should be easy to read and understand at a glance.
• **Accurate Representation:** Ensure scales, proportions, and labels accurately reflect the data. Avoid misleading visuals (e.g., truncated Y-axis on bar charts).
• **Clear Labeling:** Use informative titles, axis labels (with units!), legends, and annotations to provide context.
• **Audience Awareness:** Tailor the complexity and style of your visualization to your intended audience.
• **Color Choice:** Use color purposefully to highlight categories or trends. Be mindful of color blindness and avoid overly bright or distracting palettes.

Theory: Common Chart Types and When to Use Them 📊

Choosing the right chart is crucial for conveying your message.

1. Bar Chart:
   • **Use Case:** Comparing quantities across different categories. Great for showing counts, totals, or averages for discrete groups.
   • **Variations:** Vertical (column chart), Horizontal (often better for long category labels), Stacked, Grouped.
   • **Example:** Comparing sales figures across different product categories.
2. Line Chart:**
   • **Use Case:** Showing trends over time or a continuous sequence. Excellent for visualizing how a value changes.
   • **Variations:** Single line, Multiple lines (for comparing trends).
   • **Example:** Tracking website traffic month over month.
3. Pie Chart / Donut Chart:**
   • **Use Case:** Showing parts of a whole (proportions or percentages). Best used with a small number of categories (usually less than 5-6).
   • **Caution:** Can be hard to accurately compare slice sizes, especially if they are similar. Bar charts are often a better alternative for comparison.
   • **Example:** Displaying market share distribution among competitors.
4. Scatter Plot:**
   • **Use Case:** Visualizing the relationship or correlation between two numerical variables. Each point represents an observation.
   • **Example:** Plotting advertising spend vs. sales to see if there's a connection.
5. Histogram:**
   • **Use Case:** Showing the distribution of a single numerical variable. It groups values into bins (ranges) and plots the frequency (count) within each bin.
   • **Difference from Bar Chart:** Histograms show distribution of continuous data, bar charts compare discrete categories. The bars in a histogram typically touch.
   • **Example:** Visualizing the distribution of student test scores.
   [attachment_0](attachment)
6. Box Plot (Box-and-Whisker Plot):**
   • **Use Case:** Summarizing the distribution of a numerical variable, highlighting the median, quartiles (IQR), and potential outliers. Useful for comparing distributions across different groups.
   • **Example:** Comparing salary distributions across different job roles.

import matplotlib.pyplot as plt
import numpy as np # Often used together

# Sample Data
x = np.linspace(0, 10, 100) # 100 numbers from 0 to 10
y = np.sin(x)

# Create a plot
plt.figure(figsize=(8, 5)) # Optional: set the figure size
plt.plot(x, y, label='Sine Wave', color='blue', linestyle='--') 

# Add labels and title
plt.title('Simple Sine Wave Plot')
plt.xlabel('X-axis')
plt.ylabel('Y-axis')
plt.legend() # Show the label defined in plt.plot()
plt.grid(True)

# Display the plot (in Jupyter, this often happens automatically)
plt.show() 

import seaborn as sns
import matplotlib.pyplot as plt
import pandas as pd # Seaborn works great with Pandas

# Sample DataFrame (from previous chapter)
data = {
    'Name': ['Alice', 'Bob', 'Charlie'],
    'Age': [25, 30, 22],
    'City': ['New York', 'London', 'Paris']
}
df = pd.DataFrame(data)

# Create plots using Seaborn (often requires less setup)
plt.figure(figsize=(8, 5)) 
sns.barplot(x='Name', y='Age', data=df, palette='viridis') 
plt.title('Age Distribution (Seaborn)')
plt.show()

plt.figure(figsize=(8, 5)) 
sns.scatterplot(x='Age', y='Age', data=df) # Example scatter (using Age vs Age is silly here)
plt.title('Example Scatter Plot')
plt.show()

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

# Assume df has columns 'Category' and 'Value'
# Example Data:
data = {'Category': ['A', 'B', 'C', 'D'], 'Value': [15, 30, 22, 18]}
df_bar = pd.DataFrame(data)

plt.figure(figsize=(8, 5))
sns.barplot(x='Category', y='Value', data=df_bar, palette='coolwarm')
plt.title('Comparison of Values by Category')
plt.xlabel('Category Type')
plt.ylabel('Numeric Value')
plt.show()

# Assume df has columns 'Date' (time series) and 'Traffic'
# Example Data:
dates = pd.to_datetime(['2023-01-01', '2023-02-01', '2023-03-01', '2023-04-01'])
traffic = [1000, 1200, 1150, 1300]
df_line = pd.DataFrame({'Date': dates, 'Traffic': traffic})

plt.figure(figsize=(10, 5))
sns.lineplot(x='Date', y='Traffic', data=df_line, marker='o')
plt.title('Website Traffic Over Time')
plt.xlabel('Month')
plt.ylabel('Visitors')
plt.xticks(rotation=45) # Rotate x-axis labels if needed
plt.grid(True)
plt.show()

# Assume df has columns 'StudyHours' and 'ExamScore'
# Example Data:
study_hours = [2, 5, 1, 6, 4, 7, 3]
exam_score = [65, 85, 50, 90, 75, 95, 70]
df_scatter = pd.DataFrame({'StudyHours': study_hours, 'ExamScore': exam_score})

plt.figure(figsize=(8, 6))
sns.scatterplot(x='StudyHours', y='ExamScore', data=df_scatter, s=100) # s=size of points
plt.title('Relationship between Study Hours and Exam Score')
plt.xlabel('Hours Studied')
plt.ylabel('Exam Score (%)')
plt.show()

# Assume df has a numerical column 'Age'
# Example Data (using previous df):
data = {'Age': [25, 30, 22, 35, 28, 30, 45, 29, 31, 33]}
df_hist = pd.DataFrame(data)

plt.figure(figsize=(8, 5))
sns.histplot(data=df_hist, x='Age', bins=5, kde=True) # bins=number of bars, kde=density curve
plt.title('Distribution of Ages')
plt.xlabel('Age Group')
plt.ylabel('Frequency (Count)')
plt.show()