Analytics Workbench

Building Modern Analytics & AI Skills One Project at a Time

A reproducible, project-driven environment for learning modern analytics, data engineering, machine learning, and AI through hands-on projects.

---

Overview

Analytics Workbench is a personal learning laboratory designed to explore the modern data ecosystem by building real projects instead of completing isolated tutorials.

Rather than learning technologies in isolation, each notebook introduces new tools while solving practical analytics problems. Every project builds on previous work, reinforcing concepts through incremental, hands-on development.

The goal is to understand not only how individual tools work, but why they exist and when to use them.

---

Learning Philosophy

The guiding principle of this repository is simple:

Learn technologies by solving problems, not by memorizing APIs.

Every project begins with a business question.

Examples include:

• How much solar energy did we generate today?
• Can we forecast tomorrow's production?
• Which sensors are beginning to fail?
• How can we optimize inventory?
• Can machine learning improve forecasting accuracy?

Technology is introduced only when it helps answer those questions.

---

Technology Stack

Category	Technologies
Language	Python 3.12
Environment	uv
Version Control	Git
Notebook Environment	JupyterLab
Analytics	DuckDB
DataFrames	Pandas, Polars
Storage	Apache Parquet
Distributed Computing	Apache Spark
Machine Learning	scikit-learn
Gradient Boosting	XGBoost
Deep Learning	PyTorch
Visualization	Plotly, Matplotlib, Altair

---

Repository Structure

analytics-workbench/
│
├── README.md
├── setup.sh
├── pyproject.toml
├── uv.lock
│
├── requirements/
│   ├── core.txt
│   ├── notebook.txt
│   ├── visualization.txt
│   ├── ml.txt
│   ├── spark.txt
│   ├── utilities.txt
│   └── dev.txt
│
├── scripts/
│   └── verify_environment.py
│
├── notebooks/
│
├── projects/
│
├── data/
│   ├── raw/
│   ├── processed/
│   └── output/
│
└── docs/

---

Getting Started

Clone the repository:

git clone https://github.com/thescriptingcave/analytics-workbench.git
cd analytics-workbench

Run the automated setup:

./setup.sh

The setup script will:

• Create the Python environment
• Install all required packages
• Configure the Jupyter kernel
• Verify the installation
• Confirm that the environment is ready for development

---

Verify the Environment

At any time you can verify the installation by running:

uv run python scripts/verify_environment.py

The verification script tests the complete analytics stack, including:

• DuckDB
• NumPy
• Pandas
• Polars
• PyArrow
• scikit-learn
• XGBoost
• PyTorch
• Apache Spark

---

Launch JupyterLab

uv run jupyter lab

Select the Python (Analytics Workbench) kernel.

---

Learning Roadmap

This repository grows through a sequence of progressively more advanced notebooks.

Notebook	Topic	Status
01	Solar Analytics with DuckDB & Spark	✅ Complete
02	Pandas & Polars Fundamentals	⏳ Planned
03	Feature Engineering	⏳ Planned
04	Machine Learning with scikit-learn	⏳ Planned
05	Gradient Boosting with XGBoost	⏳ Planned
06	Deep Learning with PyTorch	⏳ Planned
07	Time-Series Forecasting	⏳ Planned
08	IoT Analytics	⏳ Planned
09	Apache Spark Transformations	⏳ Planned
10	End-to-End Analytics Pipeline	⏳ Planned

---

Architecture

Raw Data
    │
    ▼
CSV / JSON / Parquet
    │
    ▼
DuckDB
    │
    ▼
Feature Engineering
    │
    ▼
Apache Spark
    │
    ▼
Machine Learning
    │
    ▼
Visualization
    │
    ▼
Business Insights

---

Current Progress

Area	Status
Development Environment	✅
uv Workflow	✅
Git Repository	✅
DuckDB	✅
Apache Spark	✅
Pandas	⏳
Polars	⏳
Feature Engineering	⏳
scikit-learn	⏳
XGBoost	⏳
PyTorch	⏳
Airflow	⏳
dbt	⏳

---

Why This Repository Exists

There are countless tutorials explaining individual tools. Very few demonstrate how those tools work together to solve realistic analytics problems.

Analytics Workbench documents the process of building practical skills through incremental projects. Each notebook introduces a new capability while reinforcing concepts learned earlier, resulting in a growing collection of reusable examples that reflect how modern analytics systems are built.

Whether you're learning SQL, distributed computing, machine learning, or data engineering, the emphasis is always on understanding the complete workflow rather than isolated technologies.

---

Future Directions

Planned areas of exploration include:

• Advanced DuckDB Analytics
• Apache Spark Performance
• Airflow
• dbt
• Feature Engineering
• Time-Series Forecasting
• Deep Learning with PyTorch
• Solar Analytics
• IoT Analytics
• Supply Chain Analytics
• End-to-End Data Pipelines

---

License

This project is licensed under the MIT License.