Stop 05 · Data Analysis for Business AnalystsTopics · 8/10

Statistical Analysis Basics

In short

You don't need to be a statistician. But knowing when an average is lying to you, or when a difference is just noise, stops you from making a confident recommendation that's simply wrong.

Basic concepts like averages, distributions, and correlation versus causation — enough statistical literacy to avoid drawing overconfident conclusions from a small or biased dataset.

The purple blob at a chai stall looks surprised at one giant golden cup among small glass cups.

Think of it like this

A millionaire walks into a chai stall

Ten people at a chai stall earn about the same. A millionaire walks in, and the 'average income' of the group shoots up. Nobody else got richer. That's why one extreme value can make an average tell a completely false story.

Watch it happen

Mean vs median on delivery times (minutes)

1/5Seven delivery times, sorted. Most are between 12 and 18 minutes. One rider got stuck for 95.

The big ideas

What you need to know

Mean vs median

The mean is the average. The median is the middle value. With outliers or skewed data, the median is more honest.

Look at the spread

Two teams can have the same average but very different consistency. Check the range or standard deviation too.

Sample size matters

'4 out of 5 users liked it' from five people is weak evidence. Bigger samples give more trustworthy results.

Correlation vs causation

Two things moving together might share a hidden cause. Only experiments, like A/B tests, show cause clearly.

See it in code

A tiny example

excel

1=AVERAGE(B2:B8)2=MEDIAN(B2:B8)3=CORREL(B2:B30, C2:C30)

The mean and median of delivery times, and the correlation between two columns, such as discount and order value. CORREL returns a number between -1 and 1.

Save yourself the pain

Common mistakes

Oops

Reporting only the average for salaries, delivery times or order values.

Do this instead

Show the median too, and mention outliers. Skewed data makes averages misleading.

Oops

Declaring a winner after testing on 20 users.

Do this instead

Small samples swing randomly. Get more data, or say clearly that the result is early and uncertain.

Your mission

Catch a misleading average

~20 min

Quick check

Most orders are between 300 and 600 rupees, but a few bulk orders are over 50,000. Which number best shows a typical order?

Go deeper

1Calculate mean and median for a sample dataset and note when they diverge significantly.
2Learn why correlation between two metrics doesn't automatically imply one causes the other.
3Practice questioning a data-based claim for sample size or selection bias.