Knowledge Center

A concise, growing library of what Reverie can do. Use this page as both customer-facing help and as the assistant’s ground truth for supported features.

Quick Start

1. Go to the Lab, upload a CSV/XLSX. If the file has no headers, use the “Headers” panel to fix them.
2. Click summary to verify columns and see nulls/numerics/date breakdown.
3. Pick a report type (pivot/timeseries/compose) and run. Export CSV if needed.

Typical file size limit in the demo is ~50 MB. Larger, scheduled loads can use SFTP or the API.

Uploads & Header Fixes

• Header row present? Tick “File has a header row” at upload.
• Use a specific row as header (e.g., row 0) in the Headers panel of the Lab.
• Assign all columns at once: provide a list matching the file’s width.
• Rename a few: old:new lines (e.g., col_0: date).

Report Types (Canned Examples)

Don’t see what you need? Use Compose to define exactly what to group, filter, and aggregate—no new ETL required.

Compose (Ad-hoc Analyses)

Compose lets you describe table ops in JSON. The Lab includes click-to-fill examples. Structure:

{
  "type": "compose",
  "params": {
    "select": ["date","region", {"alias":"revenue","expr":"amount"}],
    "derive": [{"alias":"month","expr":"to_month(date)"}],
    "filter": [{"col":"region","op":"in","value":["EMEA","NA"]}],
    "groupby": ["month","region"],
    "aggregate": [{"col":"revenue","fn":"sum"}],
    "window": [{"alias":"mom_change","expr":"pct_change(revenue_sum)","partition":["region"],"order":["month"]}],
    "sort": [{"col":"revenue_sum","dir":"desc"}],
    "limit": 10,
    "chart": {"kind":"bar","x":"month","y":"revenue_sum","series":"region"}
  }
}

Supported helpers: to_month, to_week, to_day, year, month, day, arithmetic (A+B, A-B, A*B, A/B), filters (==, between, contains, lists), groupby + Aggs (sum, mean, count, …), window pct_change.

Scheduling & Email

Templates can target a dataset (or alias), run on a schedule (Render Cron), and email CSV/PNG outputs to a list.

• Create a template via API with a report_spec (any analysis or compose).
• Schedule the job and send to distribution (e.g., “Daily 8AM KPIs”).

Key API Endpoints

• POST /analytics/upload → dataset_id (supports has_header, header_row, assign_columns_json)
• GET /analytics/profile?dataset_id=… → shape/columns/nulls/numeric/date/top-k/preview
• POST /analytics/analyze with {"dataset_id","spec"} → table/result
• POST /datasets/assign_headers → fix headers after upload
• POST /templates/upsert, POST /jobs/run-template → scheduled emails
• POST /feedback → capture thumbs/reasons/comments

See the Lab for live examples and exports.

FAQ

What’s a “null”? A missing value in a cell. The summary shows null counts per column so you can spot data quality issues quickly.

Can I analyze files without headers? Yes—upload, then fix headers in the Lab (use a row as header or assign names).

How big can files be? Demo limit ~50 MB via browser. For bigger/recurring loads, use SFTP or the API.