Skip to content

Roadmap

The phased plan for Data Science & AI Cheat Sheets. Checkboxes track what is done and what is next.

Current release: v1.0.0 — 154 cheat sheets across 29 categories (2026-09-07).

Vision

Build the most useful, community-maintained cheat sheet collection for the data science and AI stack — comprehensive in coverage, consistent in structure, and easy to navigate. Every category starts empty and fills in through contributions, guided by priority.

Phase 0 — Repository Foundation

The structural foundation of the repository.

  • Core documentation: README, CONTRIBUTING, ROADMAP, CHANGELOG, LEARNING-PATHS
  • Central navigation index (NAVIGATION.md)
  • 30 numbered category directories with index pages
  • Reusable cheat sheet template
  • Internal link validation of all foundation documents

Phase 1 — Python & Data Foundations

The most-referenced sheets, which also unblock every learning path.

  • Python essentials — 8 sheets: basics, OOP, stdlib, regex, typing, venvs, pip, uv (01-Python)
  • pandas and NumPy — plus Polars, SciPy, and statsmodels: 5 sheets (02-Data-Analysis)
  • Data visualization — Matplotlib, seaborn, Plotly, Bokeh, Altair, folium, GeoPandas: 7 sheets (03-Data-Visualization)
  • SQL essentials — plus PostgreSQL, MySQL, SQLite, MongoDB, Redis: 6 sheets (12-SQL-Databases)
  • Git & GitHub workflows — git commands + GitHub collaboration: 2 sheets (20-Git-GitHub)
  • Mathematics & statistics — linear algebra, calculus, probability, statistics, optimization, hypothesis testing, distributions: 7 sheets (28-Mathematics-Statistics)

Phase 2 — Machine Learning & Deep Learning

Core modeling sheets for practitioners and students.

  • Machine learning core — scikit-learn, preprocessing, model selection, evaluation, feature engineering, clustering, dimensionality reduction: 7 sheets (04-Machine-Learning)
  • PyTorch essentials (05-Deep-Learning)
  • TensorFlow / Keras essentials (05-Deep-Learning)
  • Computer vision core — OpenCV, Pillow, torchvision, YOLO/Ultralytics: 5 sheets (11-Computer-Vision)
  • NLP core — NLTK, spaCy, Hugging Face transformers, datasets, tokenizers, sentence-transformers, preprocessing: 7 sheets (06-NLP)

Phase 3 — GenAI Stack

Prompting, LLMs, RAG, and agents — the fastest-growing area.

  • LLM concepts, Ollama, Hugging Face, LiteLLM: 4 sheets (08-LLM)
  • Generative AI — OpenAI, Anthropic, Gemini APIs, LangChain, LangGraph, LlamaIndex: 7 sheets (07-Generative-AI)
  • RAG architecture — embeddings, chunking, retrieval, reranking, vector DBs (FAISS, Chroma, Qdrant, Pinecone, Weaviate, Milvus), evaluation: 12 sheets (09-RAG)
  • AI agents — agent loop, tools, tool-calling, memory, planning, routing, multi-agent, evaluation: 8 sheets (10-AI-Agents)

Phase 4 — Engineering & Operations

Shipping and operating AI systems.

  • Docker and Docker Compose — containers + multi-service apps: 2 sheets (17-Docker)
  • DevOps, CI/CD, GitHub Actions, and Kubernetes — pipeline practices, workflows, env config, cluster ops: 5 sheets (16-DevOps)
  • MLOps: MLflow, DVC, deployment, monitoring, versioning — 6 sheets (15-MLOps)
  • Data engineering: Airflow, ETL/ELT, data pipelines — 3 sheets (14-Data-Engineering)
  • Big data: Spark, PySpark, Kafka, Delta Lake, Databricks — 5 sheets (13-Big-Data)
  • Cloud essentials — fundamentals, AWS, Azure, GCP: 4 sheets (21-Cloud)

Phase 5 — Maintainability & Community

Make the repository scale as contributions grow.

  • Documented repository conventions — naming, folders, headings, code blocks, tables, links, official-docs policy, version/date policy (CONTRIBUTING.md)
  • Contributor checklist and review checklist for pull requests (CONTRIBUTING.md)
  • New-technology submission process, update process, and deprecation process for obsolete tools (CONTRIBUTING.md)
  • CI pipeline that validates internal links on every PR
  • Auto-generated table of contents and search index
  • Static site / docs build from the markdown
  • Formal maintainer review process and triage labels
  • Badges and repo metadata polish

Suggesting Roadmap Items

Open an issue with the label roadmap describing the item, why it matters, and who it helps. Items that unblock multiple learning paths are prioritized first.