Stop 03 · Python for Data ScienceTopics · 3/9

Pandas for Data Manipulation

In short

If you'll learn one data library really well, make it pandas. It's how most data scientists load, filter, group and reshape tables every single day.

The standard tool for loading, exploring, filtering and transforming tabular data in Python — DataFrames, Series, groupby, merge — used in nearly every data science workflow.

The mascot sorts pages from a giant register into neat coloured piles on a desk.

Think of it like this

A college register with superpowers

Picture the office register of every student, with columns for name, branch, city and fees. Now imagine asking it 'show only CSE students from Pune, sorted by fees' and getting the answer instantly. That's a pandas DataFrame.

Watch it happen

Big orders per city, step by step

read_csv

Filter

groupby

Sort

Result

1/5read_csv loads 50,000 orders into a DataFrame with columns like city and amount.

The big ideas

What you need to know

DataFrame and Series

A DataFrame is a table. Each column is a Series with one data type and a shared index.

Filter with conditions

df[df['city'] == 'pune'] keeps matching rows. Combine conditions with & and |.

Split, apply, combine

groupby splits rows into groups, applies a calculation, then combines the results into a new table.

merge joins tables

pd.merge works like a SQL join, matching rows from two tables on a shared column.

See it in code

Walk through the code

python

1import pandas as pd2 3df = pd.read_csv("orders.csv")          # order_id, city, amount, date4print(df.head())5 6big = df[df["amount"] > 500]7by_city = (big.groupby("city")["amount"]8              .agg(["count", "mean"])9              .sort_values("count", ascending=False))10print(by_city.head(3))

1/4Load the CSV and peek at the first five rows.

Load, filter, group, sort. This four-step pattern answers a surprising number of business questions.

Save yourself the pain

Common mistakes

Oops

Looping over rows with iterrows to do maths.

Do this instead

Use column operations like df['a'] * df['b']. They're vectorised and far faster.

Oops

Changing a filtered slice and getting a SettingWithCopyWarning.

Do this instead

Use .loc[rows, 'col'] = value on the original, or call .copy() on the slice first.

Your mission

Answer three questions

~30 min

Quick check

Which line gives the average amount per city?

Go deeper

1Load a CSV into a DataFrame and explore it with .head(), .info(), .describe().
2Practice filtering, grouping, and merging DataFrames on a sample dataset.
3Learn .loc versus .iloc for label-based versus position-based selection.