1
CO2
Python Libraries & Data Visualization
NumPy, Pandas, Matplotlib, Scikit, Bokeh — and six chart types — explored through a Cricket Matches case study. Every chart draws itself.
Ten experiments — each an animated, hands-on lab manual you can run in the browser. Drag sliders, watch models fit, read the maths as it happens, and revise with built-in videos and viva questions.
NumPy, Pandas, Matplotlib, Scikit, Bokeh — and six chart types — explored through a Cricket Matches case study. Every chart draws itself.
Mean, median, mode, standard deviation & variance on the Titanic data — with a tap-to-add stats playground and a missing-value cleaner.
Ridge & Lasso. A live λ dial runs a real ridge solver — slide from overfit to underfit — plus a Ridge-vs-Lasso shrinkage demo.
Built from a CSV. Pick the weather and watch Bayes' theorem decide step by step, then run a test set for accuracy and a confusion matrix.
Two linked plots — fit + residuals. Switch data shapes (linear, curved, funnel, outlier) and let the auto-diagnosis read the pattern.
A retail case study: an RFM purchase predictor with a live decision boundary, plus a five-model leaderboard you can rank by any metric.
An interactive decision tree that traces why each applicant is approved or rejected — the explainability a bank needs to defend a decision.
Healthcare classification from medical test results. A live kernel playground draws SVM decision regions — switch Linear vs RBF and watch a ring no line can split get separated.
A manufacturing study: do three machines differ? A live ANOVA engine computes a real F-statistic and exact p-value — slide the means apart and watch the verdict flip at p = 0.05.
Build a series from trend, seasonality and noise, then forecast it forward with a confidence band that widens with the horizon — honest forecasting you can tune live.
Experiment_*.html files in the same folder as this page so the “Open lab” links work. Each lab is fully self-contained and runs offline (only fonts and the concept videos load from the web).