The Dataquest Download

Build data and AI skills – one newsletter at a time

Our Summer Flash Sale is on. Get 57% off Lifetime and unlock every current and future path, course, and project for life. Start here

Here’s what’s inside:

• Top Read: The best AI projects to build in 2026, sequenced for hiring, with the three you should actually build. Read the guide

• Webinar Recording: Catch Winning Jeopardy on demand: analyze 20,000 Jeopardy questions with Python and pandas. Watch now

• From the Community: Standout learner projects on SQL sales analysis and NYC SAT scores, plus a web-scraping resource list. Join the discussion

• Data & Resources: Five new sources: structured LLM outputs, Hugging Face datasets, BigQuery, the Census API, and the NIST AI playbook. Explore now

• What We're Reading: Fine-tuning FLAN-T5, Python 3.14's new JIT compiler, and AI's arrival in mathematics. See the picks

Top Read

Forget building another chatbot clone. We sequenced the ten best AI projects for 2026 from beginner foundations to portfolio-ready builds — each with prerequisite checks, realistic time estimates, and the four signals hiring managers actually look for: retrieval, structured output, evaluation, and deployment.

Every recommendation was shaped by the Dataquest team who built our AI Engineering in Python career path from scratch — Anna Pershyna (CTO), Anna Strahl (Curriculum Director), and Mike Levy (Content Developer).

Not sure where to start? The guide points you to the three projects worth building, beginning with a Document Q&A assistant that teaches the core RAG pattern, so your weekends go toward work that gets you hired.

Upcoming Webinar

Missed it live? Watch the walkthrough of the Winning Jeopardy project: analyze 20,000 real Jeopardy questions with Python and pandas, clean and normalize the text, and use statistics to find which terms are associated with high-value clues. Every step runs in your browser, no setup required.

From the Community

• Sales Performance and Customer Insights Analysis: Neha's project stands out for its clearly defined objective, well-explained problem context, sound methodology, efficient SQL queries, strong business impact, thoughtful discussion of intermediate observations, and insightful final findings.

• NYC Schools SAT Score Analysis: Daniel's project features an explicitly framed research question and an in-depth statistical analysis of the factors that most influence SAT scores, while also outlining the limitations of the analysis and potential directions for future work.

• Resources to Learn Web Scraping: Mamta provides a curated list of beginner-friendly books, documentation, and hands-on guides to learn both basic and advanced web scraping concepts in a practical, step-by-step manner beyond introductory tutorials

Data & Resources [New]

• Structured model outputs · OpenAI: Practice making LLM outputs reliable by forcing responses into JSON schemas for extraction, grading, and app workflows.

• Datasets documentation · Hugging Face: Load, inspect, and share NLP, audio, and computer-vision datasets, useful for model training and benchmark projects.

• BigQuery public datasets · Google Cloud: Query real hosted datasets with SQL and build cloud analytics projects without managing storage.

• Census Data API User Guide · U.S. Census Bureau: Use official demographic and economic data for API practice, joins, mapping, and public-policy analysis projects.

• NIST AI RMF Playbook · NIST AI Resource Center: Turn responsible AI into a project checklist: map risks, define measures, and document controls for AI systems.

What We're Reading

• Fine-tune FLAN-T5 for chat & dialogue summarization · shared by Anna Strahl: Working with open-source models has become much more accessible with tools like Hugging Face Transformers. This walkthrough covers dataset preparation, training, evaluation, and inference. A useful read for understanding how open-source language models can be adapted for practical NLP tasks.

• Python 3.14 and its new JIT compiler · shared by Mike Levy: Python 3.14 introduces an experimental JIT compiler, one of the language's most significant performance updates in years. This piece explains how it works, which workloads benefit, and why it matters even if you never think about Python internals.

• The AI revolution in math has arrived · shared by Brayan Opiyo: What changes when AI becomes a research collaborator rather than an answer machine? Mathematicians are making progress in days on problems that once took months, but only when an expert guides the process and checks every step.

Learning with Friends

Bring an accountability partner along on your learning journey. Refer a friend, and they’ll enjoy an extra 20% off when they subscribe, while you earn $20. It’s a win-win for everyone. Learn together, stay motivated, and use your bonuses for digital gift cards, prepaid cards, or charity donations!