Python AI Reference

Completed Jupyter notebooks from IBM/Coursera courses, organized as a personal cheat sheet. Each folder maps to a topic area — open the notebook that matches what you want to remember.

Quick lookup

I want to remember how to…	Go to
Explore data, plot distributions, find correlations	01-data-analysis/
Clean data, handle missing values, normalize features	01-data-analysis/data-wrangling-practice.ipynb
Build regression / classification pipelines	02-machine-learning/
Tune hyperparameters with GridSearchCV	02-machine-learning/practice-project-titanic.ipynb
Use decision trees, random forest, XGBoost	02-machine-learning/decision-trees.ipynb
Cluster with K-Means, DBSCAN, HDBSCAN	02-machine-learning/kmeans-customer-segmentation.ipynb
Reduce dimensions with PCA, t-SNE, UMAP	02-machine-learning/pca.ipynb
Build a full ML pipeline (weather prediction)	02-machine-learning/final-project-aus-weather.ipynb
Implement backpropagation from scratch	03-deep-learning/backpropagation.ipynb
Train models with Keras / TensorFlow	03-deep-learning/regression-with-keras.ipynb
Build CNNs for image classification	03-deep-learning/convolutional-neural-networks-keras.ipynb
Build attention, positional encoding, BERT	04-nlp-transformers/
Fine-tune Hugging Face transformers	05-generative-ai/fine-tuning-transformers-pytorch.ipynb
Build RAG (retrieval-augmented generation)	05-generative-ai/rag-pytorch.ipynb
Use DPO / PPO for LLM alignment	05-generative-ai/dpo-fine-tuning.ipynb
Build an AI agent with tools	05-generative-ai/weather-agent-daily-dish.ipynb

Folder structure

python-ai-reference/
├── 01-data-analysis/          # EDA, wrangling, model development basics
├── 02-machine-learning/       # Supervised & unsupervised ML
├── 03-deep-learning/          # Neural nets, Keras, CNNs
├── 04-nlp-transformers/       # Word2Vec, BERT, GPT, seq2seq
├── 05-generative-ai/          # LLM fine-tuning, RAG, RLHF, agents
├── requirements/              # Per-topic dependency lists
└── scripts/                   # Utilities used to organize this repo

All notebooks by topic

01 — Data Analysis

Notebook	What it covers
exploratory-data-analysis-cars.ipynb	EDA on automotive dataset
exploratory-data-analysis-laptops.ipynb	EDA on laptop pricing data
data-wrangling-review.ipynb	Data wrangling review exercises
data-wrangling-practice.ipynb	Missing data, normalization, binning, dummies
model-development-practice.ipynb	Linear & polynomial regression, pipelines

02 — Machine Learning

Notebook	What it covers
logistic-regression.ipynb	Binary classification
multi-class-classification.ipynb	Multi-class pipelines
house-sales-king-county.ipynb	Full regression project (pipelines, Ridge)
decision-trees.ipynb	Decision tree classification
decision-tree-svm-credit-fraud.ipynb	Trees + SVM for fraud detection
regression-trees-taxi-tip.ipynb	Regression trees
knn-classification.ipynb	K-nearest neighbors
random-forest-xgboost.ipynb	Ensemble methods comparison
regularization-linear-regression.ipynb	Ridge, Lasso, Elastic Net
evaluating-classification-models.ipynb	Confusion matrix, metrics
evaluating-random-forest.ipynb	Residuals, feature importances
evaluating-kmeans-clustering.ipynb	Elbow method, silhouette
kmeans-customer-segmentation.ipynb	Customer segmentation
comparing-dbscan-hdbscan.ipynb	Density-based clustering
pca.ipynb	Principal component analysis
tsne-umap.ipynb	Nonlinear dimensionality reduction
practice-project-titanic.ipynb	GridSearchCV + pipelines (Titanic)
final-project-aus-weather.ipynb	Full ML pipeline (weather prediction)

03 — Deep Learning

Notebook	What it covers
artificial-neural-networks.ipynb	ANN fundamentals
backpropagation.ipynb	Backprop from scratch
regression-with-keras.ipynb	Keras regression
convolutional-neural-networks-keras.ipynb	CNNs with Keras

04 — NLP & Transformers

Notebook	What it covers
classifying-documents.ipynb	Document classification
creating-nlp-data-loader.ipynb	Custom NLP data loaders
integrating-word2vec-part1.ipynb	Word2Vec from scratch
integrating-word2vec-part2.ipynb	Gensim Word2Vec
sequence-to-sequence-model.ipynb	Seq2seq with attention
attention-positional-encoding.ipynb	Attention mechanism
data-preparation-bert.ipynb	BERT data prep
encoder-baby-bert.ipynb	Build BERT from scratch
decoder-gpt-models.ipynb	GPT-style decoder models
transformers-classification.ipynb	Hugging Face classification

05 — Generative AI

Notebook	What it covers
loading-models-huggingface.ipynb	Load & run HF models
fine-tuning-transformers-pytorch.ipynb	Fine-tune with PyTorch
pretraining-finetuning-pytorch.ipynb	Pre-train then fine-tune
adapters-pytorch.ipynb	Parameter-efficient adapters
optional-pretraining-llms.ipynb	LLM pre-training
rag-pytorch.ipynb	RAG with PyTorch
rag-huggingface.ipynb	RAG with Hugging Face
dpo-fine-tuning.ipynb	Direct Preference Optimization
ppo-trainer.ipynb	PPO / RLHF
weather-agent-daily-dish.ipynb	AI agent with tools

Setup

Install dependencies for the topic you need:

pip install -r requirements/data-analysis.txt        # 01-data-analysis
pip install -r requirements/machine-learning.txt    # 02-machine-learning
pip install -r requirements/deep-learning.txt       # 03-deep-learning
pip install -r requirements/nlp-transformers.txt    # 04-nlp-transformers
pip install -r requirements/generative-ai.txt       # 05-generative-ai

Most notebooks download datasets automatically from IBM/Coursera S3 URLs at runtime.

Push to GitHub

cd python-ai-reference
git init
git add .
git commit -m "Organize completed Python AI notebooks as reference library"
git branch -M main
git remote add origin https://github.com/YOUR_USERNAME/python-ai-reference.git
git push -u origin main

Replace YOUR_USERNAME with your GitHub username.

Notes

• All exercise cells are filled in with working solutions.
• Duplicate notebook copies from the original folder were consolidated (e.g. only one copy of Adapters in PyTorch).
• Original files remain in the parent Generative AI stuff folder untouched.