Statistics - lesson 2 | HackIndia Skills

Key point

Shape beats one number

A dataset has a shape. The shape tells you if most values are packed together, spread out, split into groups, or pulled by a few extreme values.

Before you claim “users finish in 4 minutes” or “students spend ₹200”, look at the full distribution. Your insight should match the shape, not just the centre.

What to look for

A histogram shows how often values fall into ranges. A box plot shows the middle half, the median, and possible outliers.

For hackathon work, look for four things: skew, outliers, clusters, and range. These usually change the story.

Your to-do

Read the shape

Do each one yourself, then tap it to tick it off. The ticks are only a checklist for you: they are not marked or scored.

Check where most values sit.
Look for a long tail on the left or right.
Mark extreme values that may be real or may be data errors.
See if there are separate groups, like first-year and final-year users behaving differently.
Write one sentence on what the shape means for your feature or analysis.

0 of 5 done

Example

Plot your column

Use this on a CSV from your project. It could be response time, order value, study minutes, delivery distance, or app session length.

Ask an AI tool to help adapt the column name, but check that it did not change your data or silently remove rows.

import pandas as pd
import matplotlib.pyplot as plt

# Example: survey.csv has a column called session_minutes
df = pd.read_csv("survey.csv")

print(df["session_minutes"].describe())

plt.hist(df["session_minutes"].dropna(), bins=10)
plt.xlabel("Session length in minutes")
plt.ylabel("Number of users")
plt.title("Distribution of session length")
plt.show()

df.boxplot(column="session_minutes")
plt.ylabel("Minutes")
plt.title("Session length box plot")
plt.show()

Common mistake

Do not hide tails

A common mistake is to report only the average when the data has a long tail. If most users finish in 2 to 5 minutes but two users took 45 minutes, the average may sound too high.

Another mistake is deleting outliers because they look ugly. First ask why they exist. A 300-minute session may be a bug, an idle tab, or a real power user.

Tip

Use plain labels

Your chart title should say what is measured, not what you want to prove. Use “Distribution of checkout time” instead of “Our checkout is fast”.

In your submission, pair each chart with one honest sentence: “Most sessions are under 5 minutes, but a few long sessions pull the average up.”