foundation Estimated learning time: 8 h

3.2 Pandas: series, dataframes, indexing

You can load a messy CSV and get a clean dataframe.

Before:01. Mathematics for Machine Learning02. Python — Basics to AdvancedUnlocks:05. Classical Machine Learning

Pandas is where more hours of this course will be spent than any other library, and dataframes are the standard vehicle for tabular data. This topic builds the habits that keep sessions sane: loc and iloc used deliberately, dtypes watched, memory understood. It follows NumPy because a dataframe is arrays with labels. The named trouble is the SettingWithCopyWarning family — chained indexing sometimes modifies a copy instead of the data, and the habit of single, explicit indexing avoids the whole class of bug.

Work through these

  • Reading and writing CSV, Excel, Parquet, SQL

    Getting data in and out of the four formats you will meet most: delimited text, spreadsheets, a columnar file format, and a database. Each has its own gotchas on types and encoding.

    SQLBolt — Interactive SQL Lessons · Tutorial
  • loc, iloc, boolean masks, query

    The several ways to select rows and columns, by label, by position, by condition, or by a written expression. Mixing them up is the source of the most common confusion in this library.

  • Index and MultiIndex

    The row labels are an object in their own right, and they can have several levels. Understanding this explains a great many otherwise baffling results after grouping or reshaping.

  • dtypes, categoricals and memory use

    Column types decide both how much memory a table takes and what operations are available on it. Converting repeated text to a category type frequently cuts memory dramatically.

Sign in to keep your progress.

Free resources

Links last checked 29 Aug 2026.

Stuck here?

Ask a mentor. A real person answers, and they can see exactly which topic you're on. Usually within a couple of working days.

Checking your session…

Topics shown in module order.