Data analysis — Tools reference

You don’t call these directly — the Data analysis agent picks the right tool from your question. This is what’s available to it.

Every result is honest

Each tool returns a structured result: the method chosen, a plain‑language summary, the values, and a trust block:

"trust": { "level": "moderate",
           "caveats": ["Moderate sample (120 rows) — reasonable but not definitive.",
                       "This is a before/after comparison, not a causal test."] }

level is one of high / moderate / low / none / unassessed. When a tool declines, declined is true and no misleading number is returned.

Analytics (free, run locally)

Area	Tools
Explore	`profile`, `run_sql`, `detect_outliers`, `group_aggregate`, `compare_periods`
Customers	`rfm`, `retention_cohorts`
Association	`analyze_association`, `association_matrix`, `market_basket`
Prediction	`train_classifier`, `train_regressor`, `feature_importance`, `explain_prediction`
Time series	`forecast`, `decompose`, `detect_changepoints`
Causal	`causal_effect` (declines unless the data can support a credible estimate)
Clustering	`cluster`, `profile_clusters`, `reduce_dimensions`

All analytics are free forever and run entirely on your machine.

Account & reports (pro / trial)

Tool	Purpose
`account_status`	Show your role (`free` or `pro`).
`save_report`	Save a report to your account.
`list_reports`, `get_report`	Browse saved reports.
`create_folder`, `list_folders`	Organize reports.

Saving reports needs an active trial or pro subscription — see Pricing. Saved reports land on your dashboard, where you can organise and share them.

Table Intelligence — built by Shubham Randive. Questions? Get in touch.