200 exercises to build real Pandas intuition, on real data.
Most people learn Pandas by reading the documentation for one method at a time.
That teaches you what groupby does. It doesn't teach you to look at a messy
CSV and recognize that this is a groupby problem — and that recognition is
most of the job.
Pandas Workout is 200 exercises designed to build that instinct. Every one
uses a real public dataset, with the ordinary problems real data has: missing
values, inconsistent types, dates stored as strings, numbers stored with
currency symbols. Every one comes with a full solution that explains the
reasoning, not just the code.
What you'll work through
- Cleaning data that arrived in the wrong shape
- Rows, columns, and the indexes that tie them together
- Reading and writing CSV, Excel, JSON and SQL
- Text manipulation and the
.straccessor - Dates, times, and resampling
- Grouping and aggregate calculations
- Plotting, when a number isn't enough
440 pages, from Manning Publications.
Get the book
- Manning — print and ebook, and includes access to liveBook
- Amazon
- O'Reilly — included with a Learning Platform subscription
Manning Publications, April 2024. ISBN 9781617299728.
The book ends. Bamboo Weekly doesn't.
Two hundred exercises sounds like a lot until you finish them. Bamboo Weekly is
what comes next: a new Pandas exercise every week, built the same way the
book's are — a real public dataset, a genuine question someone might actually
ask of it, and a full worked solution that explains the thinking.
It has been running since 2022, and it's free.
Read this week's issue · Browse exercises by skill · Subscribe
About the author
I'm Reuven M. Lerner, and I've taught Python and Pandas full time since 1995 —
to engineers at Apple, Cisco, IBM, Intel, PayPal, VMWare, Western Digital and many others. I wrote Pandas Workout and Python Workout for Manning, I
publish Bamboo Weekly, and I teach Python and data-analysis courses at
LernerPython.com.