📊 PackIA: Data Science
From dirty CSV to model and dashboard with agents
Direct answer: From dirty CSV to model and dashboard with agents This pack brings together 7 resources for Claude, Claude Code, ChatGPT Codex.
Agentic notebooks that clean, explore, model, and visualize: pandas, scikit-learn, and applied statistics with assumption checks and interpretable explanations for the business.
Included in the Completo — $50
Plans are cumulative: the Pro Plan includes all Essential Plan packs, and the Complete Plan includes all packs (Essential + Pro). This pack unlocks with the Completo ($50).
Buy with the Completo ($50) Or build your custom plan
For ChatGPT Codex: this pack is used from the Codex terminal, where MCP servers, subagents, prompts, and guides support coding and lookup tasks. View for ChatGPT Codex: only the resource types compatible with ChatGPT Codex (MCP Server, Agent, Prompt collection, Guide). The full pack has 7 resources.
Use cases
- A complete EDA with business insights and publishable charts
- A predictive model with cross-validation and explained metrics
- A reproducible data pipeline with quality tests
Master prompt: Senior Data Scientist
You are a data scientist. Dataset: {dataset} with columns {columns} and target {goal}. 1) EDA: nulls, outliers, correlations, and 5 business insights, 2) justified feature engineering, 3) a baseline + 2 models with validation appropriate for {problem_type}, a comparison table, and a justified choice, 4) business-aligned metrics and a decision threshold, 5) clean, reproducible, production-ready commented code.
Included resources (5)
- duckdb / motherduck (MCP Server, motherduckdb/mcp-server-motherduck, ★ 521 · official) — Analytical SQL queries over DuckDB files and databases from the assistant. GitHub
- TaskWeaver (Agent, microsoft/TaskWeaver, ★ 6.2k · official) — Code-first agents that plan and execute data analysis by generating and verifying code. GitHub
- postgres-mcp (pro) (MCP Server, crystaldba/postgres-mcp, ★ 3.3k) — Postgres MCP Pro: read and write access to PostgreSQL, performance analysis, and index tuning. GitHub
- LangGraph (Agent, langchain-ai/langgraph, ★ 41.8k) — Low-level orchestration for stateful, long-running agents, with deterministic or dynamic flows. GitHub
- Prompt Engineering Guide (Prompt collection, dair-ai/Prompt-Engineering-Guide, ★ 78.4k) — Guide, papers, and notebooks on prompt engineering, RAG, and agents: few-shot, CoT, and self-consistency. GitHub