Data Analysis + LLM Curriculum (Python, R, PydanticAI with Ollama)

A 12-week, project-based roadmap for learning data analysis, Unix/Linux basics, and LLM apps using Python, R, PydanticAI, and Ollama (local models).
All resources are 100% free and include lectures, interactive projects, GitHub practice, and an optional "cool kids" track with Neovim + CLI tools.

---

Week -1: Mac + Homebrew Setup Checklist

Step 1: Install Python + Tools

• Install Python 3.11 → brew install [email protected]
• Install uv package manager → brew install uv
• Install Git → brew install git
• Install iTerm2 → brew install --cask iterm2
• Install VSCode → brew install --cask visual-studio-code
• Verify Python version → python3 --version

Step 2: Install R + RStudio

• Install R → brew install --cask r
• Install RStudio → brew install --cask rstudio

Step 3: Install JupyterLab

• Install JupyterLab → uv pip install jupyterlab
• Run Jupyter → jupyter lab

Step 4: Install Data Science Libraries

• Install NumPy → uv pip install numpy
• Install pandas → uv pip install pandas
• Install matplotlib → uv pip install matplotlib
• Install seaborn → uv pip install seaborn
• Install SciPy → uv pip install scipy
• Install statsmodels → uv pip install statsmodels

Step 5: Install AI Stack

• Install Pydantic → uv pip install pydantic
• Install PydanticAI → uv pip install pydantic-ai
• Install ChromaDB → uv pip install chromadb

Step 6: Install Ollama

• Install Ollama → brew install ollama
• Pull Llama3.1 model → ollama pull llama3.1
• Pull nomic-embed-text model → ollama pull nomic-embed-text
• Test Ollama → ollama run llama3.1 "Hello world"

Step 7: Create Project Folder

• Create project folder → mkdir data-analyst-curriculum && cd data-analyst-curriculum
• Create virtual environment → uv venv .venv
• Activate environment → source .venv/bin/activate

---

Suggested Repo Layout

data-analyst-curriculum
├── 01_unix_linux
├── 02_python_basics
├── 03_pandas_eda
├── 04_stats_intro
├── 05_r_tidyverse
├── 06_pydanticai_rag
├── 07_pydanticai_assistant
├── 08_sql_tooling
├── 09_eval_observability
└── capstones
    ├── analyst_copilot_pydanticai
    └── r_eda_report

---

Week 0: CS & Unix/Linux

Lectures

• CS50’s Introduction to Computer Science (Harvard)
• freeCodeCamp – Linux for Beginners (4 hrs)
• MIT Missing Semester – Shell Tools & Scripting

Interactive

• OverTheWire – Bandit Wargame (hands-on Linux terminal practice)

---

Weeks 1–3: Python Foundations + Unix/Linux Basics

Lectures

• freeCodeCamp – Python for Beginners (4 hrs)
• Kaggle – Python Course
• CS50’s Introduction to Python (Harvard)

Interactive Projects

• freeCodeCamp – Scientific Computing with Python Certification
  Projects: Arithmetic Formatter, Time Calculator, Budget App, Polygon Area Calculator, Probability Calculator

---

Weeks 4–5: Statistics & R

Lectures

• Khan Academy – Statistics & Probability
• StatQuest with Josh Starmer
• MIT OCW – Probability & Statistics
• freeCodeCamp – R Programming for Data Science (2 hrs)
• R for Data Science (Book)

Interactive Projects

• freeCodeCamp – Data Analysis with Python Certification
  Projects: Mean-Variance Calculator, Demographic Data Analyzer, Medical Data Visualizer, Time Series Visualizer, Sea Level Predictor

---

Weeks 6–7: Python for Data Analysis

Lectures

• freeCodeCamp – Data Analysis with Python (5 hrs)
• Kaggle – Pandas Course
• Kaggle – Data Visualization Course

Interactive

• Finish freeCodeCamp Data Analysis with Python Certification

---

Weeks 6–12: LLMs, PydanticAI & AI Agents

Lectures

• freeCodeCamp – LangChain for Beginners (concepts apply to PydanticAI)
• DeepLearning.AI – Free Short Courses on LLMs

---

Week 10: SQL & Databases

Lectures

• freeCodeCamp – SQL Full Course (4 hrs)
• Khan Academy – Intro to SQL

Interactive Projects

• freeCodeCamp – Relational Database Certification
  Projects: Celestial Bodies Database, World Cup Database, Salon Scheduler, Number Guessing Game

---

Weeks 11–12: Capstones & Projects

Resources

• Kaggle – Free Datasets
• DataCamp Workspace (Free Tier)
• MIT OCW – Data Science & Machine Learning

Interactive Projects

• freeCodeCamp – Machine Learning with Python Certification
  Projects: Rock, Paper, Scissors AI, Cat/Dog Classifier, Book Recommender, Stock Predictor, Neural Network SMS Classifier

---

Git & GitHub (Do This Throughout)

• GitHub – Hello World Guide
• freeCodeCamp – Git & GitHub for Beginners (2 hrs)
• Oh My Git! – Interactive Git Game

Weekly GitHub Habits

• Create a repo for each week/project
• Add README.md summarizing what was learned
• Commit often (git add . && git commit -m "message" && git push)
• Upload Jupyter Notebooks (render well on GitHub)
• Pin best repos to showcase skills

---

⚡ Optional: "Cool Kids" Neovim + CLI Tools Track

Neovim Setup

• Install Neovim → brew install neovim
• Learn basics: i (insert), :wq (save & quit), dd (delete line), / (search)
• freeCodeCamp – Vim Tutorial for Beginners (2 hrs)
• Explore config: ~/.config/nvim/init.lua

CLI Tools (Make Linux fun & powerful)

• Install fzf (fuzzy finder) → brew install fzf
• Install ripgrep (fast search) → brew install ripgrep
• Install htop (system monitor) → brew install htop
• Install bat (better cat) → brew install bat
• Install exa (better ls) → brew install exa
• Bonus: Try tmux for terminal multiplexing → brew install tmux

Practice

• Replace basic commands: cat → bat, ls → exa
• Use fzf to search command history
• Use ripgrep to search code fast
• Manage processes with htop
• Keep multiple projects open in tmux

---

✅ Finish Week -1 through Week 12, sprinkle in the optional “cool kids” track, and she’ll have a full data + AI + Linux toolkit with a strong GitHub portfolio.

📊 Data Analysis + LLM Projects (Week-by-Week)

This project roadmap matches the 12-week curriculum. All projects are aligned to tools introduced that week, and many include LLM integration using PydanticAI and Ollama.

---

✅ Week 0: CS & Unix/Linux

Project: Command Line File Audit Tool

• Write a Python CLI tool to summarize number, type, and size of files in a directory.
• Optional: Output as a CSV report.

---

✅ Week 1: Python Fundamentals

Project: Python Expense Tracker (CLI)

• Build a simple CLI that logs expenses and categorizes them.
• Output to CSV with pandas.
• Stretch: Add terminal charts with rich or plotext.

---

✅ Week 2: Python Functions + CLI Skills

Project: CSV Column Analyzer

• Let users input a CSV and return column types, null counts, and stats.
• Use pydantic-ai to summarize or rephrase the report via LLM.

---

✅ Week 3: Basic EDA in Python

Project: Netflix Dataset EDA

• Use pandas, matplotlib, and seaborn to explore the Netflix dataset.
• Ask Ollama to suggest plots or describe patterns with pydantic-ai schema outputs.

---

✅ Week 4: Statistics with Python

Project: Survey Summary Generator

• Take a small dataset of survey results.
• Use Python to calculate means, modes, distributions.
• Ask LLM to generate a Markdown report with PydanticAI.

---

✅ Week 5: Intro to R & Tidyverse

Project: R Data Explorer

• Use R (dplyr, ggplot2) to generate a basic report of any dataset.
• Export tables/plots.
• Bonus: Knit an .Rmd to HTML.

---

✅ Week 6: Advanced Pandas

Project: Airbnb Dataset EDA Bot

• Use Kaggle's NYC Airbnb dataset.

• Build a CLI or notebook that allows user to:

  • Load data
  • Ask questions
  • Get answers powered by Ollama + PydanticAI

---

✅ Week 7: Data Cleaning with LLMs

Project: Dirty CSV Fixer

• Input: Messy CSV (with typos, missing units, bad formats).
• Define a schema with pydantic-ai.
• LLM suggests cleaned rows and explanations.

---

✅ Week 8: CSV Q&A Assistant

Project: CSV Chat Agent

• Upload a CSV and ask natural questions ("What’s the avg revenue in Q1?").
• Use pydantic-ai + chromadb + ollama for retrieval.

---

✅ Week 9: PDF Report Analyzer

Project: PDF → Markdown Extractor

• Load a multi-page PDF report (e.g. economic data).
• Use PyMuPDF to extract text.
• Ask LLM to generate a summary and key metrics as Markdown.

---

✅ Week 10: SQL + LLM

Project: Natural Language to SQL

• Load a SQLite database (e.g. ecommerce).
• User types: "Top 5 customers by revenue."
• LLM generates SQL → fetch results → show as table or chart.

---

✅ Week 11: Capstone Prep

Project: Analyst Copilot (Part 1)

• Combine multiple tools:

  • Ask a question
  • Pull from CSV + SQL
  • Output: Graphs + Summary + raw data

• Use pydantic-ai to structure answers.

---

✅ Week 12: Capstone Final

Project: Analyst Copilot (Final)

• Wrap up with a CLI or Web UI (Streamlit).
• Multi-modal assistant: PDF + CSV + SQL.
• Structured answers + human-friendly output.
• Host on GitHub with full documentation.

---

These projects ensure hands-on practice across data analysis, LLM-powered workflows, and reproducible portfolio-building with Python, R, SQL, and PydanticAI.