Unit 10 • Lesson 1

Choosing a Chart

Overview

The shape of the question decides the shape of the picture.

Intermediate 20–25 min

← Previous Next: Labels and Honesty

What You Will Learn in This Lesson

By the end of this lesson, you will be able to:

Match four common question shapes to four chart types.
Say what the table behind each chart looks like.
Explain why a pie chart is usually the wrong answer.
Build the summary table before thinking about the picture.

Why This Matters

Choosing the chart is not a decoration step — it is the last analysis decision. A bar chart and a line chart of the same numbers make different claims, and picking the one that matches your question is what stops the picture saying something the data does not.

Step 1: Four questions, four charts

Almost every chart you need is one of these, and the question tells you which.

The mapping

Comparison across categories is a bar. Change over time is a line. The distribution of one column is a histogram. The relationship between two numbers is a scatter.

# Comparison  -> bar        mean score per team
# Over time   -> line       sales per month
# Distribution-> histogram  how scores are spread
# Relationship-> scatter    hours against score
print("the question decides, not the tool")

Why bars work

Bars share a baseline and encode value as length, which is the comparison people read most accurately. A pie encodes it as angle, which they read worst — and it can only ever show parts of one whole.

import pandas as pd

df = pd.DataFrame({"team": ["red", "red", "blue"], "score": [90, 80, 70]})
print(df.groupby("team")["score"].mean().round(1))

Mini Practice #1: Name the chart

Try It Yourself

For each question, write down which of the four charts it asks for before reading on.

# 1. How does revenue change across the twelve months?
# 2. Which of our five regions sells most per head?
# 3. Are longer study hours associated with higher marks?
# 4. What do the marks look like -- clustered, or spread out?

print("line, bar, scatter, histogram")

Press Run to see output

Step 2: What the code looks like

The chart is one call. The table behind it is the part worth getting right, and the part you can check.

Table first

Every chart in this lesson is a summary table plus a .plot. If the table is wrong, the chart is wrong and much more convincing.

import pandas as pd

df = pd.DataFrame({"team": ["red", "red", "blue"], "score": [90, 80, 70]})
summary = df.groupby("team")["score"].mean().round(1)
print(summary)
# summary.plot(kind="bar")   # the picture, once the table is right

Sort a bar chart by value

Alphabetical order carries no information. Ordering the bars by the thing being compared is what makes the ranking readable at a glance.

import pandas as pd

df = pd.DataFrame({"team": ["zeta", "alpha", "mu"], "score": [70.0, 90.0, 80.0]})
means = df.groupby("team")["score"].mean()
print(means)
print(means.sort_values(ascending=False))

The exception is a category with a natural order — months, age bands, school years. Sorting those by value destroys the thing the reader is using to navigate.

Mini Practice #2: Order the bars

Try It Yourself

Produce the mean per team and then the same table ordered so the ranking is readable.

import pandas as pd

df = pd.DataFrame({"team": ["zeta", "alpha", "mu"], "score": [70.0, 90.0, 80.0]})

means = df.groupby("team")["score"].mean()
print(means)
print(None)   # ordered by value

Press Run to see output

Step 3: When to Use It

Comparing categories

A bar chart of one summary per group, sorted by value.

df.groupby("team")["score"].mean().sort_values()

Showing change over time

A line chart of a series with dates on the index.

When Not to Use It

Do not use a pie chart to compare more than two or three parts, or a line chart to connect categories that have no order.

End-of-Lesson Exercises

Exercise 1: The table behind a bar chart

Write chart_data(df) that returns the table behind a bar chart of mean score per team, as a dictionary of team to mean rounded to one decimal place.

A bar chart of "mean by team" is a group-by with the means rounded for display. Get that table right and the chart is one more line.

import pandas as pd

def chart_data(df):
    # {team: mean score} -- the table a bar chart would draw.
    return {}

print(chart_data(pd.DataFrame({"team": ["red"], "score": [90]})))

Press Run to try it, then Check my work to mark it.

Exercise 2: Bars in a readable order

Write ranked_bars(df) returning a list of (team, mean) pairs ordered highest mean first, with the means to one decimal place. Alphabetical order carries no information.

Group to a mean per team, round, sort_values(ascending=False), then build plain (name, value) tuples with str() and float().

import pandas as pd

def ranked_bars(df):
    # [(team, mean)] highest first.
    return []

print(ranked_bars(pd.DataFrame({"team": ["zeta", "alpha"], "score": [70.0, 90.0]})))

Press Run to try it, then Check my work to mark it.

← Prev Next