37 concepts worth remembering.
Each experience has four depth layers. Start with intuition, manipulate the mechanism, open technical detail only when useful, then jump into the full reference.
Python: How Code Becomes State
What actually changes when Python executes one line?
Functions & Reusable Reasoning
Why do functions make programs easier to trust?
Choosing Python Collections
List, tuple, set or dictionary—what changes when you choose the wrong one?
pandas & Table Thinking
What does a DataFrame represent beyond “an Excel-like table”?
SQL Query Flow
How does SQL turn tables into evidence?
Excel as an Analytical Model
When is a spreadsheet a model rather than just a grid?
Missing Data Decisions
A blank cell is not a method—what should you decide before imputing?
Scaling & Distance
Why can a harmless-looking unit change alter a model?
Categorical Encoding
How do categories become numbers without inventing false meaning?
Exploratory Data Analysis
What should you learn before fitting a model?
Correlation Analysis
How do direction, strength, outliers and nonlinearity change correlation—and when can correlation help select features?
Visualisation & Storytelling
What makes a chart explain rather than decorate?
Train, Validation & Test
Why do we need data the model has never seen?
Cross-Validation
How do we estimate generalisation when one split is too fragile?
Data Leakage
Can a model look excellent because we accidentally gave it the answer?
Linear Regression
What does “best-fitting line” really optimise?
Logistic Regression
How does a linear model become a probability classifier?
Decision Trees
How does a tree convert questions into regions of feature space?
Random Forest
Why can many unstable trees become a stable model?
Gradient Boosting
How can a sequence of weak models repair its own mistakes?
K-Nearest Neighbours
What happens when prediction is literally “look at similar examples”?
Support Vector Machines
Why do only a few boundary points determine the classifier?
K-Means Clustering
How do unlabeled points organise themselves around centroids?
Principal Component Analysis
How can we rotate a feature space without throwing away its main variation?
Classification Metrics
Why can “90% accuracy” be terrible?
Regression Metrics & Residuals
What does one error number hide?
Bias, Variance & Learning Curves
Is poor validation performance caused by underfitting or instability?
Regularisation
How do we tell a flexible model not to chase every detail?
Hyperparameter Optimisation
How do we search a large configuration space without fooling ourselves?
Probability Calibration & Thresholds
Is a predicted 0.8 probability actually an 80% event rate?
Explainability & Model Interpretation
How do p-values, feature AUC, permutation importance and RFE tell different stories?
Predictive Uncertainty
When should a model say “I am not sure”?
Time-Series Forecasting
Why can’t time-ordered data be shuffled like ordinary rows?
Neural Networks & Backpropagation
How does a network turn many simple transformations into complex behaviour?
Attention & Transformers
How can each token decide which other tokens matter?
Graph Neural Networks
How can information move across relationships rather than rows?
Deployment, Drift & Monitoring
What changes after a model leaves the notebook?