transformar a business question em dados discovery, SQL, e a plain-English answer. — Claude Skill
Um Skill Claude para Claude Code por Astronomer — executar /analyzing-data no Claude·Atualizado em 18 de jun. de 2026·vmain@789b454
Encontra conceitos e tabelas certos no data warehouse, prepara a query de análise, valida pressupostos e devolve uma resposta com ressalvas compreensíveis para negócio.
- Translates stakeholder questions em métricas, dimensions, filters, e dados fontes.
- Discovers tables e prior analysis patterns antes writing novo SQL.
- Returns o answer, caveats, e recommended próximo analysis in business language.
- Caches useful learnings so repeated questions get easier over time.
A stakeholder asks para a number, but métrica definition e fonte table pressupostos stay hidden.
Run /analyzing-data para make definitions, fonte selection, query logic, result, e caveats visible.
Para quem é
O que faz
Answer ativação, retenção, conversão, ou usage questions a partir de warehouse dados.
Analyze pipeline, win taxa, segment performance, ou campanha contribution.
encontrar o right fonte tables e definitions antes building a dashboard.
Como funciona
Clarify o decisão, métrica, segment, date range, e acceptable caveats.
Look up relevant concepts, tables, prior queries, e known definitions.
Draft SQL ou analysis steps that match o métrica definition.
validar results against simple reasonableness verifica.
Explain o result, caveats, e suggested follow-up.
Opções de entrada
o decisão o analysis deve suporte.
Exemplo
Question: Did o onboarding checklist improve ativação? Definition: ativação means workspace invited 2+ teammates e created primeiro project dentro de 7 days. lançar date: June 1. Segments: SMB, mid-market, enterprise. Known tables: product_events, workspaces, contas. precisar de: analysis approach, result summary, caveats, e próximo analysis.
métrica: 7-day ativação. comparar workspaces created antes June 1 com workspaces created depois June 1. Segment by conta size. Exclude workspaces created less than 7 days ago because they have não had enough time para activate.
ativação appears para improve depois lançar in SMB e mid-market. Enterprise moves less clearly because sample size is smaller e vendas-assisted onboarding may follow a different path.
| Segment | antes lançar | depois lançar | Change | Readout | |---|---:|---:|---:|---| | SMB | 36% | 44% | +8 pts | Likely positive movement | | Mid-market | 42% | 51% | +9 pts | Strong sinal para investigate | | Enterprise | 37% | 42% | +5 pts | Directional only | | todos | 38% | 46% | +8 pts | Improvement, não proof de causality |
This is a antes/after analysis, não an experiment. verificar aquisição mix, seasonality, evento tracking stability, e whether workspaces had a full 7-day window.
Break down checklist completion by step. If invite step completion rose, o checklist likely helped. If only traffic mix changed, do não credit o checklist yet.
Métricas que melhora
Funciona com
Quer usar Análise de Dados?
Escolha como começar.
Instale e execute este skill localmente no seu computador.
Abra um terminal no seu computador e cole este comando:
Isto descarrega o skill com todos os ficheiros para o seu computador:
Adicione -g no fim para o tornar disponível em todos os seus projetos.
Inicie o Claude Code, depois escreva o comando:
dados Analysis
Answer business questions by querying o dados warehouse. o kernel auto-starts on primeiro ¤KEEP0¤ call.
todos CLI commands below are relative para this skill's directory. antes running qualquer ¤KEEP0¤ command, ¤KEEP1¤ para o directory containing this file.
workflow
- Pattern lookup — verificar para a cached query estratégia:
uv run scripts/cli.py pattern lookup "<utilizador's question>"
If a pattern exists, follow its estratégia. Record o resultado depois executing:
```bash
uv run scripts/cli.py pattern record <nomear> --success # ou --failure
- Concept lookup — encontrar known table mappings:
uv run scripts/cli.py concept lookup <concept>
3. **Table discovery** — If cache misses, pesquisar o codebase (¤KEEP0¤) ou query ¤KEEP1¤. See [reference/discovery-warehouse.md](reference/discovery-warehouse.md).
4. **Execute query**:
```bash
uv run scripts/cli.py exec "df = run_sql('SELECT...')"
uv run scripts/cli.py exec "print(df)"
- Cache learnings — Always cache antes presenting results:
Cache concept → table mapping
uv run scripts/cli.py concept learn <concept> <TABLE> -k <KEY_COL>
Cache query estratégia (if discovery was needed)
uv run scripts/cli.py pattern learn <nomear> -q "question" -s "step" -t "TABLE" -g "gotcha"
6. **Present findings** para utilizador.
## Kernel Functions
| Function | Returns |
|----------|---------|
| ¤KEEP0¤ | Polars DataFrame |
| ¤KEEP0¤ | Pandas DataFrame |
¤KEEP0¤ (Polars) e ¤KEEP1¤ (Pandas) are pre-imported.
## CLI Reference
### Kernel
```bash
uv run scripts/cli.py warehouse list # List warehouses
uv run scripts/cli.py start [-w nomear] # Start kernel (com opcional warehouse)
uv run scripts/cli.py exec "..." # Execute Python code
uv run scripts/cli.py status # Kernel status
uv run scripts/cli.py restart # Restart kernel
uv run scripts/cli.py stop # Stop kernel
uv run scripts/cli.py install <pkg> # Install package
Concept Cache
uv run scripts/cli.py concept lookup <nomear> # Look up
uv run scripts/cli.py concept learn <nomear> <TABLE> -k <KEY_COL> # Learn
uv run scripts/cli.py concept list # List todos
uv run scripts/cli.py concept import -p /path/to/warehouse.md # Bulk import
Pattern Cache
uv run scripts/cli.py pattern lookup "question" # Look up
uv run scripts/cli.py pattern learn <nomear> -q "..." -s "..." -t "TABLE" -g "gotcha" # Learn
uv run scripts/cli.py pattern record <nomear> --success # Record resultado
uv run scripts/cli.py pattern list # List todos
uv run scripts/cli.py pattern delete <nomear> # Delete
Table Schema Cache
uv run scripts/cli.py table lookup <TABLE> # Look up schema
uv run scripts/cli.py table cache <TABLE> -c '[...]' # Cache schema
uv run scripts/cli.py table list # List cached
uv run scripts/cli.py table delete <TABLE> # Delete
Cache Management
uv run scripts/cli.py cache status # Stats
uv run scripts/cli.py cache claro [--stale-only] # claro
References
- reference/discovery-warehouse.md — Large table handling, warehouse exploration, INFORMATION_SCHEMA queries
- reference/common-patterns.md — SQL templates para tendências, comparisons, top-N, distributions, cohorts
Documentos de referência
name: analyzing-data description: Queries dados warehouse e answers business questions about dados. Handles questions requiring database/warehouse queries including "who usa X", "how many Y", "show me Z", "encontrar clientes", "what is o count", dados lookups, métricas, tendências, ou SQL analysis.
dados Analysis
Answer business questions by querying o dados warehouse. o kernel auto-starts on primeiro ¤KEEP0¤ call.
todos CLI commands below are relative para this skill's directory. antes running qualquer ¤KEEP0¤ command, ¤KEEP1¤ para o directory containing this file.
workflow
- Pattern lookup — verificar para a cached query estratégia:
uv run scripts/cli.py pattern lookup "<utilizador's question>"
If a pattern exists, follow its estratégia. Record o resultado depois executing:
```bash
uv run scripts/cli.py pattern record <nomear> --success # ou --failure
- Concept lookup — encontrar known table mappings:
uv run scripts/cli.py concept lookup <concept>
3. **Table discovery** — If cache misses, pesquisar o codebase (¤KEEP0¤) ou query ¤KEEP1¤. See [reference/discovery-warehouse.md](reference/discovery-warehouse.md).
4. **Execute query**:
```bash
uv run scripts/cli.py exec "df = run_sql('SELECT...')"
uv run scripts/cli.py exec "print(df)"
- Cache learnings — Always cache antes presenting results:
Cache concept → table mapping
uv run scripts/cli.py concept learn <concept> <TABLE> -k <KEY_COL>
Cache query estratégia (if discovery was needed)
uv run scripts/cli.py pattern learn <nomear> -q "question" -s "step" -t "TABLE" -g "gotcha"
6. **Present findings** para utilizador.
## Kernel Functions
| Function | Returns |
|----------|---------|
| ¤KEEP0¤ | Polars DataFrame |
| ¤KEEP0¤ | Pandas DataFrame |
¤KEEP0¤ (Polars) e ¤KEEP1¤ (Pandas) are pre-imported.
## CLI Reference
### Kernel
```bash
uv run scripts/cli.py warehouse list # List warehouses
uv run scripts/cli.py start [-w nomear] # Start kernel (com opcional warehouse)
uv run scripts/cli.py exec "..." # Execute Python code
uv run scripts/cli.py status # Kernel status
uv run scripts/cli.py restart # Restart kernel
uv run scripts/cli.py stop # Stop kernel
uv run scripts/cli.py install <pkg> # Install package
Concept Cache
uv run scripts/cli.py concept lookup <nomear> # Look up
uv run scripts/cli.py concept learn <nomear> <TABLE> -k <KEY_COL> # Learn
uv run scripts/cli.py concept list # List todos
uv run scripts/cli.py concept import -p /path/to/warehouse.md # Bulk import
Pattern Cache
uv run scripts/cli.py pattern lookup "question" # Look up
uv run scripts/cli.py pattern learn <nomear> -q "..." -s "..." -t "TABLE" -g "gotcha" # Learn
uv run scripts/cli.py pattern record <nomear> --success # Record resultado
uv run scripts/cli.py pattern list # List todos
uv run scripts/cli.py pattern delete <nomear> # Delete
Table Schema Cache
uv run scripts/cli.py table lookup <TABLE> # Look up schema
uv run scripts/cli.py table cache <TABLE> -c '[...]' # Cache schema
uv run scripts/cli.py table list # List cached
uv run scripts/cli.py table delete <TABLE> # Delete
Cache Management
uv run scripts/cli.py cache status # Stats
uv run scripts/cli.py cache claro [--stale-only] # claro
References
- reference/discovery-warehouse.md — Large table handling, warehouse exploration, INFORMATION_SCHEMA queries
- reference/common-patterns.md — SQL templates para tendências, comparisons, top-N, distributions, cohorts