Lab work
Laboratory exercises for the course — Labs 1–4 are available now, the rest are coming soon. Start with the connected lecture, predict what will happen, then run the notebook and compare your output with your prediction.
Connecting two tables, one step at a time
Connect two tables without changing what a row means.
6 guided sections · About 2 hours · Jupyter notebook0/6 sections reviewedOpen labWhat an average hides
Use your own case to calculate, plot and explain the data.
8 guided sections · About 2 hours · Jupyter notebook0/8 sections reviewedOpen labData Cleaning and Quality
Make supported cleaning choices and explain what remains uncertain.
6 guided sections · 110 minutes + 10-minute check · Jupyter notebook0/6 sections reviewedOpen labTransformations, Pipelines and Prediction
Follow the supplied split from raw records to an honest test result.
5 guided sections · PDF + worked solution0/5 sections reviewedOpen labComing soon 9 labs
Colab and Python
Python for Data Science — for students new to programming
Exploratory analysis that survives scrutiny
Structured exploration; confounding; Simpson's paradox
Chart critique and redesign
Visual encoding; baselines; colour and accessibility; message titles
Probability and simulation
Probability; conditional direction; Bayes with natural frequencies
Intervals, resampling and power
Sampling variability; confidence intervals; tests; effect size; A/B design
The regression workflow
Baselines; linear regression; metrics; residuals; cross-validation
Classification, end to end
Logistic regression; k-NN and trees; confusion matrix; thresholds; fairness
Clustering with honest uncertainty
Distance; k-means; choosing k; PCA; naming responsibly
Time series, text and a dashboard
Time series and text; dashboards; privacy; contestability
How to work through a lab
- Open a section and review its lecture links.
- Read the task and write your prediction before running code.
- Download the notebook, open it in Jupyter or upload it to Google Colab, and run the cells in order.
- Use your own output to answer the questions. Record the counts, units and evidence.
- Return here to review the connection and save a short learning note.
Code on this site is for reading — it does not execute Python or generate your personalised results. Downloaded study copies have saved outputs cleared; Lab 3 identity fields are reset to placeholders. Lab 4 uses a PDF and a guided worked example.