Understand

Beginner in data · after basic Python

Data Science & Analytics

Beta This course is a beta version. Lessons, exercises and translations are still being reviewed and may change. Report anything that looks wrong with the feedback button.

A data scientist's job is not to produce charts. It is to answer a question and say how far the answer can be trusted. You'll frame questions, clean data and document what you changed, simulate the statistics until they make sense, query databases, and communicate limits as clearly as results. Exercises run and are checked in your browser; nothing to install until the capstone.

Modules
14
Lessons
39
Time
~29 hours
Projects
3
Get started

What you'll be able to do

  • Turn a vague request into a question data could actually answer
  • Clean a messy table with pandas, and document every transformation
  • Explain sampling variability, confidence intervals, p-values, power and multiple testing by simulating them
  • Separate correlation from causation, and analyse a randomised experiment correctly
  • Write analytical SQL with joins, common table expressions and window functions
  • Choose charts that don't mislead, and show uncertainty
  • Deliver a reproducible investigation with its limits stated

Who it's for

People who can write Python functions and loops and want to work with data properly, including those who were taught statistics as a list of formulas.

What you need

Python functions, loops, lists and dictionaries (the first levels of the Python course). School algebra. No statistics needed. Every chapter is open, so skip what you already know. pandas downloads about 10 MB the first time an exercise uses it.

What you'll build

  • The Delivery Delays
  • Bike-Sharing Demand Analysis
  • Independent Data Investigation

Curriculum

Thinking with DataQuestions, evidence and the tools for numerical work. 3 modules · 8 lessons
  1. 1 The Data Science MindsetQuestions before code, provenance before charts.
  2. 2 Thinking in ArraysNumPy, shapes, vectorisation and the linear algebra machine learning actually uses.
  3. 3 Numbers in MemoryBroadcasting, dtypes and floating-point traps.
Wranglingpandas, joins, messy records and data-quality reports. 2 modules · 8 lessons
  1. 4 pandas and TablesSelect, group, join and reshape, without losing rows by accident.
  2. 5 Cleaning Data, and Saying SoMissing values, invalid records, outliers, duplicates, time, and reports.
Inference and ExperimentsWhat samples can tell you about the world, and when causes can be claimed. 3 modules · 8 lessons
  1. 6 Describing and SamplingSummaries, probability, and why samples vary.
  2. 7 Estimation and TestingConfidence intervals, p-values, power, and the ways they get abused.
  3. 8 Experiments and CausesConfounding, Simpson's paradox and A/B tests.
Professional PracticeVisualisation, SQL, exploration and communication. 4 modules · 9 lessons
  1. 9 Visualising HonestlyEncodings, uncertainty, and charts that don't lie.
  2. 10 SQL for AnalyticsJoins, CTEs, window functions and data-quality queries.
  3. 11 Exploratory Data AnalysisProfile, segment, investigate, and keep exploration apart from confirmation.
  4. 12 Professional AnalyticsReproducible reports, privacy, and telling people what they need to hear.
InvestigationsCase studies with less and less guidance, then your own capstone. 2 modules · 6 lessons
  1. 13 Case StudiesLess guidance each time.
  2. 14 CapstoneYour own investigation, from question to defensible conclusion.