Explore. Explain. Enable better decisions.
DataForScience is a home for practical data science: thoughtful analysis, dependable software, and open learning resources. We build tools and share workflows that help turn complex data into evidence people can use.
Our work values the full path from a well-framed question to a result others can inspect, reproduce, and extend.
| Area | Focus |
|---|---|
| 🔎 Data analysis | Finding meaningful patterns through exploratory analysis, statistics, and visualization. |
| 🤖 Machine learning | Building interpretable, well-evaluated models for real-world decisions. |
| 🧰 Data engineering | Creating reliable pipelines, clean datasets, and reusable analytical foundations. |
| 📚 Open learning | Sharing notebooks, examples, and guides that make data science more approachable. |
Question → Data → Exploration → Validation → Insight → Action
- Start with the problem. Useful analysis begins with a clear decision or question.
- Make work reproducible. Code, environments, and assumptions should be easy to revisit.
- Communicate with clarity. A result is only valuable when people can understand its context and limits.
- Build for reuse. Small, well-documented tools compound into durable capability.
The organization brings together hands-on code, notebooks, slides, and reusable tools. These projects are a good place to start; browse the full repository collection to explore more.
- Probability and Statistics for Everyone — Hands-on foundations in descriptive statistics, probability, Bayesian analysis, model fitting, and hypothesis testing.
- Causal Inference for Everyone — Practical work on association, intervention, counterfactuals, A/B testing, and causal questions.
- Transforming Excel Analysis into Python and pandas — A bridge from spreadsheet analysis to reproducible, programmable data workflows.
- Data Visualization with Matplotlib and Seaborn — Principles and practical examples for turning information into clear visual stories.
- Network and Graph Algorithms from Scratch — Core graph structures and algorithms behind ranking, recommendation, graph databases, and web crawling.
- NLP with PyTorch — Natural-language processing foundations, embeddings, neural networks, and real-data examples.
- Large Language Models for Data Science — Using LLMs for summarization, translation, entity recognition, data processing, and prompt design.
- GraphRAG — A hands-on pipeline from raw text to a knowledge-graph-backed chatbot with auditable retrieval.
- CrewAI for Production-Ready Multi-Agent Systems — Agent teams, custom tools, delegation, long-term memory, human review, and maintainable configuration.
- Timeseries for Everyone — Time-series workflows with pandas, ARIMA forecasting, ARCH models, and real and synthetic data.
- Deep Learning from Scratch — A practical introduction to neural-network concepts and implementations.
- Blockchain Data Science — A Python framework for templated SQL, BigQuery blockchain datasets, caching, and analytics.
- Epidemic Modeling for Everyone — Notebooks and simulations for visualizing and modeling infectious-disease dynamics.
Good contributions come in many forms: improving documentation, reporting an issue, sharing a reproducible example, or proposing a feature. Before opening a pull request, please review the repository's contribution guidance and keep changes focused, tested, and clearly described.