Core Concepts

Workspaces, data sources, pipelines, executions and charts.

Five objects form the whole platform. Understanding how they relate removes most of the guesswork from day-to-day use.

Workspaces and membership

A workspace (internally a tenant) is the isolation boundary. It owns data sources, pipelines, executions and settings. Users join a workspace through a membership record that carries a role.

Roles

Field	Type	Description
admin	string	Full control, including members, plan and destructive actions.
editor	string	Creates and edits data sources, pipelines and charts.
analyst	string	Builds charts and reports, and runs existing pipelines.
viewer	string	Read-only access to dashboards and reports.

Data sources

A data source describes where raw data comes from. Three kinds are supported today.

Source types

Field	Type	Description
csv	file	Delimited text. Header row is used for column names.
json	file	An array of objects, or a single object, flattened into columns.
api	endpoint	A REST endpoint polled at run time, with an optional bearer token.

Pipelines and transformations

A pipeline is the recipe: one data source plus an ordered list of transformations and a schedule. Transformations run top to bottom, and each step sees the output of the previous one.

Transformation types

Field	Type	Description
filter	operator	Keeps rows matching a comparison, for example amount > 100.
rename	string	Renames a column from its source name to a target name.
convert_type	string	Casts a column to string, number, decimal, boolean or JSON.
sort	string	Orders rows ascending or descending by a column.

Executions

An execution is one run of a pipeline. It records status, timings, duration, total records, success rate and a full log.

Execution statuses

Field	Type	Description
pending	string	Queued and waiting for a worker.
running	string	Actively processing; logs stream live.
success	string	Finished cleanly. A dataset was written.
failed	string	Aborted. Inspect the log for the failing step.
cancelled	string	Stopped before completion.

Datasets and charts

Only the latest successful execution produces the dataset that Analytics reads. A failed run never replaces a good dataset, so your dashboards stay available while you debug.

Charts are definitions — field choices, ordering, row limits and width — stored per pipeline. Because the dataset is columnar, filtering and sorting happen instantly in the browser.

Scheduling and frequency

Frequency options

Field	Type	Description
hourly	string	Runs at the top of every hour.
daily	string	Runs once per day at the configured run time.
weekly	string	Runs on the same weekday and time.
monthly	string	Runs on the same day of each month.

Object lifecycle

From file to chart

data source  ──►  pipeline (source + transformations + schedule)
                       │
                       ├── execution #1  failed   ──► log only
                       ├── execution #2  success  ──► Parquet dataset
                       │                                   │
                       └── execution #3  success  ──► replaces dataset
                                                           │
                                                           ▼
                                             charts  ──►  dashboards

Deleting an execution removes its Parquet file and its log. Deleting a pipeline removes every execution beneath it. Deleting a data source is blocked while pipelines still reference it.

Last reviewed October 2026.