Exploratory Data Analysis (EDA) on a Public Dataset
Part of the Data Analyst Roadmap
In short
Before you can answer questions about a dataset, you need to know what's in it. An EDA project shows you can explore unfamiliar data and spot what's interesting, odd or broken.
A structured first look at an unfamiliar dataset — shape, types, distributions, missing data, and early patterns — demonstrating your process before jumping to conclusions.
Think of it like this
First visit to a new sabzi mandi
At an unfamiliar market, you don't buy from the first stall. You walk around, check what's fresh, which stalls are busy, what looks overpriced and which tomatoes are bruised. EDA is that first walk through a dataset before you commit to any answer.
Watch it happen
An EDA notebook's sections
1/5Check rows, columns and data types. Is a date stored as text? Is price a number?
The big ideas
What you need to know
Be curious, not conclusive
EDA is about noticing. Write "interesting: weekend orders are larger" and test it properly later.
Report data quality
Missing values, duplicates and odd codes are findings too. Hiring managers like analysts who spot them.
Plot before you summarise
An average can hide two very different groups. A histogram shows it instantly.
Narrate in markdown
Explain each step in plain words between code cells, so a reviewer can follow your thinking.
See it in code
Walk through the code
1import pandas as pd2 3df = pd.read_csv("listings.csv")4print(df.shape)5df.info()6print(df.isna().mean().sort_values(ascending=False).head())7print(df.describe())8print(df["city"].value_counts().head(10))9df["price"].hist(bins=50)1/4Load the file, then check its size and each column's type and non-null count.
The first cells of almost every EDA notebook. In a minute you know size, types, gaps, ranges and the biggest groups.
Save yourself the pain
Common mistakes
Oops
Dumping 40 charts with no explanation and calling it EDA.
Do this instead
Keep the charts that taught you something and write one line on what each shows.
Oops
Picking a famous, overused dataset that every portfolio has.
Do this instead
Choose something you care about, like local weather, cricket stats or your city's open data.
Your mission
First milestone: the first ten cells
Quick check
During EDA, the average delivery time is 32 minutes but the histogram shows two humps, around 20 and 50 minutes. What does that suggest?
Go deeper
- 1Choose a public dataset in a domain you find genuinely interesting.
- 2Structure the notebook: load, inspect, clean, explore, summarize findings.
- 3Resist the urge to skip straight to pretty charts before understanding the data's quirks.
More projects at this stop
End-to-End SQL + BI Dashboard Project
A/B Test Analysis Case Study
Business Metrics Deep-Dive Report
Data Cleaning Case Study (Messy Dataset)
Python + Pandas Analysis Notebook
Portfolio Website or GitHub Showcase
Stuck on this topic? Ask an Insider
Get 1:1 guidance from people who've walked this exact path — on the InsideEdge app.
