Skip to content
LumenaryAcademy

Lumenary CS · Electives

Data Science

A Lumenary CS elective: the practice of turning data into decisions — the workflow, wrangling and exploring data, visualisation, machine learning on tables, experiments in products, models in production, and doing it responsibly.

This elective of Lumenary CS teaches data science as it is practised: asking a good question, getting and cleaning the data, exploring it, modelling it, testing ideas with experiments, putting models into production, and communicating honestly what the data can and cannot say. The methods underneath are taught in Lumenary Statistics; this course puts them to work, and links there rather than re-deriving them.

It assumes Lumenary CS's Start and Core levels and its learning from data lesson, and the Foundations of Lumenary Statistics. Code is in Python with pandas, NumPy and scikit-learn, and every block with a Run in browser button runs there, with nothing to install.

It is written to inform, not to sell: tools and companies are named fairly and by link, and the lessons are as plain about costs, failures and trade-offs as about what works. Every lesson was drafted with AI assistance and is awaiting review by a human expert. The text, exercises and code are original and licensed CC BY-SA 4.0 (code Apache-2.0). Papers, specifications and projects' own documentation are linked, not retold.

AI-draftedAwaiting review by a data scientist
  • Chloe · The Statistician
    “What question are we really asking?”
  • Margaret · The Simulator
    “What happens if we try it ten thousand times?”
  • Gideon · The Sceptic
    “What does this assume, and could it be chance?”

Foundations · 4 lessons

Working with data

The daily work of a data scientist: the workflow from question to decision, wrangling data with pandas, exploratory analysis, and visualisation and storytelling.

  1. 1The data science workflowFrom a question to a decision — framing, finding and cleaning data, exploring, modelling, communicating and deciding, and keeping every step reproducible. 75 min
  2. 2Wrangling data with pandasSeries and DataFrames, reading messy files, types, dates and missing values, joins that do not lie, reshaping, group-by — and how to make it fast. 80 min
  3. 3Exploratory data analysisLook before you model — summaries and distributions, outliers and data errors, relationships and segments, and the data that make summary statistics lie. 75 min
  4. 4Visualisation and storytellingChoosing a chart for the question, what the eye judges well and badly, colour for everyone, honest axes and uncertainty, and writing the story the numbers support. 75 min

Practice · 4 lessons

Data science in practice

From analysis to impact: machine learning on tabular data, experiments in products, text, images and other kinds of data, and models in production.

  1. 5Machine learning for tabular dataPipelines, features and the four model families that matter on tables — and how to evaluate them without fooling yourself. 85 min
  2. 6Experiments in productsA/B testing as companies practise it — choosing what to measure, randomising, computing power, checking the plumbing, reducing variance, resisting the urge to peek, and experimenting on people responsibly. 85 min
  3. 7Text, images and other kinds of dataTurning words, pixels, sounds, places, networks and timestamps into numbers a model can use — and the traps particular to each. 85 min
  4. 8Data science in productionFrom a notebook to a service that keeps working — packaging and serving models, watching for drift, retraining, feature stores, the tooling, decision support, and knowing when not to deploy. 85 min

Frontier · 3 lessons

Data science in 2026

Where the field is going: working with AI assistants and agents, responsible data science, and case studies from many fields and countries.

  1. 9Data science with AI assistants and agentsLanguage models that write and run analysis code — what they do well, how they fail quietly, how to check their work, and how to stay accountable for it. 80 min
  2. 10Responsible data scienceFairness you can measure and the trade-offs you cannot escape, privacy beyond removing names, consent and minimisation, documentation, and why so much applied work fails to reproduce. 85 min
  3. 11Case studiesFive projects from public health, agriculture, sport, energy and public services on four continents — the question, the data, the method, what went right and what went wrong — with synthetic re-creations you can run. 80 min

Further reading and resources