01

Exploratory data analysis

Palmer Penguins — load, inspect, missing values, plots.

Preview — PDF (exported with outputs). Click to expand.

Open PDF in a new tab if the embed does not load.

02

Linear regression

Auto MPG — train/test split, metrics, coefficients.

Preview — PDF (exported with outputs). Click to expand.

Open PDF in a new tab if the embed does not load.

03

Logistic regression

Iris — scaling, confusion matrix, decision boundary.

Preview — PDF (exported with outputs). Click to expand.

Open PDF in a new tab if the embed does not load.

04

Decision trees

Wine dataset — train a tree, plot, feature importance.

Preview — PDF (exported with outputs). Click to expand.

Open PDF in a new tab if the embed does not load.

05

Random Forest

Sonar — rock vs mine classification.

Preview — PDF (exported with outputs). Click to expand.

Open PDF in a new tab if the embed does not load.

06

XGBoost

Iris — DMatrix API and sklearn classifier.

Preview — PDF (exported with outputs). Click to expand.

Open PDF in a new tab if the embed does not load.

07

Gradient boosting compared

XGBoost, LightGBM, CatBoost on bank marketing data.

Preview — PDF (exported with outputs). Click to expand.

Open PDF in a new tab if the embed does not load.

08

K-means clustering

Mall customers — elbow method, segments, silhouette.

Preview — PDF (exported with outputs). Click to expand.

Open PDF in a new tab if the embed does not load.

09

Principal component analysis

Penguins & wine — variance, 2D projection, reconstruction.

Preview — PDF (exported with outputs). Click to expand.

Open PDF in a new tab if the embed does not load.

10

Prophet forecasting

Microsoft stock — seasonality, future horizon, MAE.

Preview — PDF (exported with outputs). Click to expand.

Open PDF in a new tab if the embed does not load.

2.1

Intro to PyTorch

Tensors, autograd, and training a model with PyTorch.

Preview — PDF (exported with outputs). Click to expand.

Open PDF in a new tab if the embed does not load.

2.2

Simple neural network

Build and train a feedforward network in PyTorch.

Preview — PDF (exported with outputs). Click to expand.

Open PDF in a new tab if the embed does not load.

2.3

Regularization & optimization

Dropout, batch norm, and learning rate schedulers.

Preview — PDF (exported with outputs). Click to expand.

Open PDF in a new tab if the embed does not load.

2.4

CIFAR-10 CNN

Convolutional neural network for image classification.

Preview — PDF (exported with outputs). Click to expand.

Open PDF in a new tab if the embed does not load.

2.5

Sales forecasting (LSTM)

LSTM for sequence-based time-series prediction.

Preview — PDF (exported with outputs). Click to expand.

Open PDF in a new tab if the embed does not load.