Data Visualization with Matplotlib

# CHAPTER 18

Data Visualization with Matplotlib #

1. Chapter Introduction #

2. Matplotlib Basics and Pyplot #

Matplotlib is massive, but 95% of the time, data scientists only use a specific module within it called pyplot.

Installation & Importing:

!pip install matplotlib

# 'plt' is the standard alias
import matplotlib.pyplot as plt

3. The Line Chart (For Trends over Time) #

Line charts are perfect for showing time-series data (e.g., stock prices, monthly revenue).

import matplotlib.pyplot as plt

# Data
months = ['Jan', 'Feb', 'Mar', 'Apr', 'May']
revenue = [5000, 5200, 4800, 6000, 6500]

# 1. Create the plot
plt.plot(months, revenue, color='blue', marker='o') # 'o' adds dots at data points

# 2. Add labels and title (CRITICAL for good charts!)
plt.title("Monthly Revenue 2023")
plt.xlabel("Month")
plt.ylabel("Revenue in USD")
plt.grid(True) # Adds a background grid

# 3. Display the chart
plt.show()

4. The Bar Chart (For Comparing Categories) #

Bar charts are used to compare categorical data (e.g., Sales by Region, Population by Country).

regions = ['North', 'South', 'East', 'West']
sales = [350, 420, 200, 500]

plt.bar(regions, sales, color=['red', 'blue', 'green', 'orange'])
plt.title("Sales by Region")
plt.xlabel("Region")
plt.ylabel("Units Sold")
plt.show()

5. The Scatter Plot (For Finding Correlations) #

Scatter plots are used to visualize the relationship (correlation) between two numerical variables. For example, does spending more on ads increase sales?

ad_spend = [100, 200, 300, 400, 500]
sales_generated = [150, 300, 280, 550, 600]

plt.scatter(ad_spend, sales_generated, color='purple')
plt.title("Ad Spend vs Sales")
plt.xlabel("Ad Spend ($)")
plt.ylabel("Sales Generated ($)")
plt.show()

*(If the dots form a line going up and to the right, you have a positive correlation!)*

6. The Histogram (For Data Distribution) #

Histograms look like bar charts, but they serve a different purpose. They show the *distribution* of a single numerical variable. They tell you where most of your data "clumps" together.

import numpy as np

# Generate 1000 random test scores centered around 75
scores = np.random.normal(75, 10, 1000)

# bins=20 means divide the data into 20 groups (bars)
plt.hist(scores, bins=20, color='skyblue', edgecolor='black')
plt.title("Distribution of Test Scores")
plt.xlabel("Score")
plt.ylabel("Frequency (Number of Students)")
plt.show()

7. Mini Project: Revenue Analytics Dashboard #

We can combine multiple charts into one image using plt.subplots().

# Create a figure with 1 row, 2 columns of charts. Size is 12x5 inches.
fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(12, 5))

# Chart 1: Line Chart on the left (ax1)
ax1.plot(months, revenue, marker='o', color='green')
ax1.set_title("Revenue Trend")
ax1.set_ylabel("USD")

# Chart 2: Bar Chart on the right (ax2)
ax2.bar(regions, sales, color='teal')
ax2.set_title("Regional Sales")
ax2.set_ylabel("Units")

# Adjust spacing so they don't overlap, then show
plt.tight_layout()
plt.show()

8. Common Mistakes #

• Forgetting plt.show(): Depending on your environment, if you don't call plt.show(), the chart might not render, or you'll see ugly memory address text above the image.

• Naked Charts: Creating a chart without a Title, X-label, and Y-label is a cardinal sin in data science. A chart without context is just abstract art.

• Saving blank images: If you want to export a chart using plt.savefig('chart.png'), you MUST call it *before* plt.show(). plt.show() clears the canvas after displaying it.

9. MCQs #

10. Interview Questions #

• Q: A stakeholder wants to know if there is a relationship between a customer's Age and their Total Spend. What type of chart would you build and why?

• Q: Explain the difference between a Bar Chart and a Histogram.

11. Summary #

12. Next Chapter Recommendation #