Build a clean sales summary from raw transaction data using pandas — the kind of output you'd produce before a weekly review meeting.
A practical reference for the most common cleaning operations in pandas — the work that happens before every real analysis.
Call a free LLM API to classify review sentiment and extract themes — turning raw text into a structured, tagged dataset.
Translate plain-English questions into runnable SQL using the OpenAI API — a practical tool for analyst workflows.
Extract qualification signals from raw CRM notes and rank your pipeline automatically using a free LLM API.
Set up a real PostgreSQL database on Supabase's free tier, write data from Python, and query it back with filters.
Write real SQL against a live in-memory database inside Colab — no credentials, no cloud account, just Python's built-in sqlite3.
Join a three-table schema and aggregate the results by tier, category, and product — the operations behind most real business reporting.
Use RANK(), LAG(), and SUM() OVER to analyze trends across periods — plus CTEs to keep complex queries readable.
Run SQL queries against a CSV file without loading it into memory first — then add window functions with QUALIFY for ranking in one pass.
Feed Claude your column names and sample values — get a full data dictionary back in seconds, ready to publish or hand off.
Tag customers with their cohort month, compute months-since-acquisition, and pivot into the retention matrix every SaaS analyst needs to know.
Traverse any parent-child structure — org charts, category trees, bill of materials — using the recursive CTE pattern most SQL tutorials skip.
Load a DataFrame, send it to Gemini, and ask anything — get a first-pass analysis in seconds instead of writing queries from scratch.
Compute z-scores in five lines and flag statistical outliers globally and per-group — find the weird rows before your stakeholders do.