Day 3 — Data Visualization

matplotlib · CSV files · Bar charts · Scatter plots

Schedule

Download a Dataset

Pick one. If you scraped/fetched data yesterday you can use that too — ask your instructor how to load it.

Save the CSV file to the same folder as your Python script.

Install

pip install matplotlib pandas

Quick Reference — New Things Today

Reading a CSV with pandas

import pandas as pd   # pd is the shorthand everyone uses for pandas

# read_csv() loads the file into a DataFrame — think of it like a spreadsheet in Python
df = pd.read_csv("pokemon.csv")
print(df.head())                  # show the first 5 rows so you can see the structure
print(df.columns.tolist())        # list the column names exactly as they appear in the file

Accessing columns

# Get an entire column as a Python list using the column name in square brackets
names = df["Name"].tolist()
hp    = df["HP"].tolist()

# Filter rows — keep only the rows where a condition is True
high_hp = df[df["HP"] > 80]           # only Pokémon with HP greater than 80
print(high_hp[["Name", "HP"]])         # show just those two columns

Sorting

# sort_values() sorts the whole DataFrame by a column
# ascending=False means highest value first (descending order)
# .head(10) keeps only the first 10 rows after sorting
top10 = df.sort_values("Attack", ascending=False).head(10)

Making Charts

Bar Chart

plt.bar() takes two required arguments:

• First argument — x labels: a list of the text labels that appear along the bottom of the chart, one per bar. Here we use the Pokémon names.
• Second argument — heights: a list of numbers that determine how tall each bar is. The first label is paired with the first height, the second with the second, and so on — the two lists must be the same length.
• color: optional — sets the color of all bars. You can use color names like "tomato", "steelblue", "gold", or hex codes like "#ff5733".

import pandas as pd
import matplotlib.pyplot as plt   # plt is the standard shorthand everyone uses

df = pd.read_csv("pokemon.csv")

# Sort by Attack highest-first, then keep only the top 10 rows
top10 = df.sort_values("Attack", ascending=False).head(10)

plt.figure(figsize=(10, 5))   # create a figure: 10 inches wide, 5 inches tall

# plt.bar(x_labels, heights, color=...)
#   top10["Name"]   — list of Pokémon names, one label per bar (x-axis)
#   top10["Attack"] — list of Attack values, one number per bar (bar height)
#   The 1st name is paired with the 1st Attack value, 2nd with 2nd, etc.
plt.bar(top10["Name"], top10["Attack"], color="tomato")

plt.title("Top 10 Pokémon by Attack")   # title displayed above the chart
plt.xlabel("Pokémon")                   # label for the horizontal axis
plt.ylabel("Attack")                    # label for the vertical axis
plt.xticks(rotation=45, ha="right")     # rotate x labels 45° so they don't overlap
plt.tight_layout()                      # fix spacing so labels aren't cut off
plt.show()                              # open the chart window

Scatter Plot

import pandas as pd
import matplotlib.pyplot as plt

df = pd.read_csv("pokemon.csv")

plt.figure(figsize=(8, 6))
# scatter() plots one dot per row — x position is Attack, y position is Defense
# alpha=0.5 makes dots 50% transparent so overlapping dots are still visible
plt.scatter(df["Attack"], df["Defense"], alpha=0.5, color="steelblue")
plt.title("Attack vs Defense")
plt.xlabel("Attack")
plt.ylabel("Defense")
plt.tight_layout()
plt.show()

Horizontal Bar Chart

plt.barh() works just like plt.bar() but the bars go sideways instead of up. The "h" stands for horizontal. The arguments swap roles:

• First argument — y labels: a list of labels that appear along the left side of the chart (the y-axis), one per bar. Long text like movie titles fit much better here than squeezed along the bottom.
• Second argument — widths: a list of numbers that determine how far each bar extends to the right. Larger number = longer bar.
• color: same as plt.bar() — optional color for all bars.

import pandas as pd
import matplotlib.pyplot as plt

df = pd.read_csv("movies.csv")

# Sort ascending=True (lowest first), then .tail(8) takes the LAST 8 rows — the highest values
# This makes the biggest bars appear at the TOP of the horizontal chart
top8 = df.sort_values("Gross_Billions", ascending=True).tail(8)

plt.figure(figsize=(8, 5))

# plt.barh(y_labels, widths, color=...)
#   top8["Title"]         — list of movie titles, one label per bar (y-axis, left side)
#   top8["Gross_Billions"] — list of box office numbers, one per bar (bar length going right)
plt.barh(top8["Title"], top8["Gross_Billions"], color="gold")

plt.title("Top 8 Highest-Grossing Movies")
plt.xlabel("Box Office (Billions $)")   # label on the horizontal axis (the numbers)
plt.tight_layout()
plt.show()

More Chart Options — Pick Any

Every chart follows the same recipe: get two things that line up (labels + numbers, or two number lists), hand them to a plotting function, add a title and axis labels, then plt.show(). Use any dataset — the ones above, or your own data from Day 2.

Starter — one clear chart

counts = df["Type"].value_counts()          # how many of each type
plt.pie(counts, labels=counts.index, autopct="%1.0f%%")   # autopct shows percentages
plt.title("Pokémon by Type")
plt.show()

Standard — show a distribution or relationship

plt.hist(df["Points"], bins=10, color="steelblue")   # bins = number of bars
plt.title("Distribution of Points")
plt.xlabel("Points per game")
plt.ylabel("How many players")
plt.show()

Boss Challenge — multiple views at once