Unit 10 • Lesson 4

A Summary Table

Overview

The small table that goes under the chart.

Intermediate 20–25 min

← Previous Next: Writing Up a Finding

What You Will Learn in This Lesson

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

Build a table with one row per thing and one column per fact.
Order it by the thing being compared.
Include the group size.
Say what the table does that the chart cannot.

Why This Matters

A chart shows the shape; the table pins the values. Together they answer both "what is the pattern" and "what exactly was the number", and the group size in the table is what stops a reader over-reading a bar built from two rows.

Step 1: One row per thing, one column per fact

The table under a chart is small on purpose: the few numbers the picture cannot state exactly, and the counts behind them.

Count and mean together

One group-by gives both. The count is not optional — it is what makes the mean interpretable.

import pandas as pd

df = pd.DataFrame({"team": ["a", "a", "b"], "score": [90.0, 80.0, 95.0]})
print(df.groupby("team")["score"].agg(["count", "mean"]).round(1))

Name the columns for a reader

"count" and "mean" are how it was computed. "students" and "mean score" are what it means.

import pandas as pd

df = pd.DataFrame({"team": ["a", "a", "b"], "score": [90.0, 80.0, 95.0]})
out = df.groupby("team").agg(
    students=("score", "count"),
    mean_score=("score", "mean"),
).round(1)
print(out)

Mini Practice #1: Count and mean, named

Try It Yourself

Produce a summary with the columns named students and mean_score.

import pandas as pd

df = pd.DataFrame({"team": ["a", "a", "b"], "score": [90.0, 80.0, 95.0]})

print(df.groupby("team").agg(
    students=("score", "count"),
    mean_score=("score", "mean"),
).round(1))

Press Run to see output

Step 2: Order it by the thing being compared

The first row should be the one the reader is looking for. Alphabetical order almost never is.

Sort by the value

Best first, so the ranking is visible without reading every row.

import pandas as pd

df = pd.DataFrame({"team": ["a", "a", "b"], "score": [90.0, 80.0, 95.0]})
out = df.groupby("team")["score"].agg(["count", "mean"]).round(1)
print(out.sort_values("mean", ascending=False))

And match the chart

If the bars are ordered by value, the table should be too. A reader moving between the two should not have to re-find each row.

import pandas as pd

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

Mini Practice #2: Best first

Try It Yourself

Sort the same summary so the highest mean is the first row.

import pandas as pd

df = pd.DataFrame({"team": ["a", "a", "b"], "score": [90.0, 80.0, 95.0]})

out = df.groupby("team")["score"].agg(["count", "mean"]).round(1)
print(None)   # sorted by mean, best first

Press Run to see output

Step 3: When to Use It

Under a chart

The exact values the picture only shows approximately, with the group sizes.

df.groupby("team").agg(students=("score", "count"), mean_score=("score", "mean"))

When readers need to look things up

A short, sorted table answers “what exactly was it for my group”.

When Not to Use It

Do not paste the whole grouped output. A summary table that needs scrolling is data, not a summary — keep the few columns the question needs.

End-of-Lesson Exercises

Exercise 1: The table under the chart

Write summary_table(df) returning a list of (team, count, mean) tuples, highest mean first, with the mean to one decimal place.

Group once and ask for count and mean together, sort by the mean descending, then build plain tuples so the result is easy to compare.

import pandas as pd

def summary_table(df):
    # [(team, count, mean)] sorted by mean, best first.
    return []

print(summary_table(pd.DataFrame({"team": ["a"], "score": [90.0]})))

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

Exercise 2: Groups too small to compare

Write thin_groups(df, least) returning the names of the teams with fewer than least rows, best mean first. These are the bars a reader should be warned about.

Group for count and mean, keep the rows whose count is below the threshold, sort what is left by mean descending, and return the index as a list.

import pandas as pd

def thin_groups(df, least):
    # Teams with fewer than `least` rows behind them.
    return []

print(thin_groups(pd.DataFrame({"team": ["a", "a", "b"], "score": [1.0, 2.0, 9.0]}), 2))

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

← Prev Next