Data analysis with pandas

Real data is late, incomplete and spread over five formats. This course follows the whole path, from data capture to the final chart: loading, cleaning, crossing and summarising with pandas, then making the numbers talk in a graph anyone can read.

Duration: 3 days · Level: intermediate

For: analysts, statisticians, scientists and developers who need to explore and report on data.

Prerequisites: knowing the basics of Python: types, loops, functions.

Pricing Contact
Data analysis with pandas

Day 1 — Loading and cleaning

  • Jupyter and friends: a comfortable lab
  • Series and DataFrame: the two building blocks
  • Loading CSV, Excel, JSON and SQL
  • Looking at your data before trusting it
  • Selecting with loc, iloc and boolean masks
  • Fixing types and encodings
  • Missing data: find it, fill it or drop it
  • Duplicates and outliers
  • Text columns with the str accessor

Day 2 — Making data talk

  • Dates, periods and time series
  • Resampling and rolling windows
  • Grouping and aggregating with groupby
  • Crossing sources with merge and join
  • Reshaping: pivot, melt, stack
  • Categorical data
  • Applying your own functions, and why to avoid it
  • Method chaining for readable analyses
  • Descriptive statistics

Day 3 — Showing the results

  • Charts with matplotlib
  • Statistical plots with seaborn
  • Interactive charts with plotly
  • Choosing the right chart for the message
  • Exporting to Excel, CSV and parquet
  • From notebook to reproducible script
  • Speeding up pandas: vectorisation and types
  • A word about polars and duckdb
  • Workshop: from data capture to the final graph

Elevate Your Python Proficiency!

Contact Examples of programs