analytics

Dispatch here when the task is to extract insight from data. The team analyzes datasets and produces findings — statistical summaries, models, or visualizations that communicate what the data shows. It does not collect or move data (that is data-engineering) and does not produce written narrative beyond what is needed to communicate the findings.

---

analytics-lead

The analytics-lead frames the analytical question, decides which analysis is needed, dispatches to the data-scientist or data-visualizer as appropriate, reviews findings for correctness and clarity, and delivers a summary with supporting artifacts. It requests clarification when the question is ambiguous or the data source is unspecified, and declares completion when the analytical question is answered with evidence.

Tools: standard workgroup-lead tools Skills: digest

---

data-scientist

Dispatch when the task requires statistical analysis, modeling, inference, or pattern detection. Works with structured data and produces findings, summaries, and model outputs. Not for producing charts or dashboards (data-visualizer) or for moving and transforming data between systems (data-engineering).

Tools: Bash, Read, Write, Glob Skills: python, r-statistical-analysis, digest

---

data-visualizer

Dispatch when findings need to be communicated through charts, graphs, or dashboards. Works from analysis results or structured datasets; does not perform statistical analysis. Specify the target medium — static image, interactive, or embedded in a document.

Tools: Bash, Write, Read Skills: charting, digest