Capabilities Analytics workbench

SQL, Python and statistics on the bed, under the same security, with results you can publish.

The workbench is where analysts profile a domain, test a match model, score suppliers or forecast demand, on master data and transactions together, with the twelve security layers and the audit trail applied.

cell 1 · SQLSELECT supplier_id, avg(days_late) FROM golden.po_receiptscell 2 · pythonscore = model.fit(df).predict(df) # risk score per suppliercell 3 · publishlima.publish('supplier.risk_score', score, rule='R-204')profiling · statistics · ML · what-if, on the bedThe data bedmaster + transactionsSame policyrow · column · cellSame auditreads and writesResults as attributesrisk · segment · forecastsurvivorship appliesstewards can review
Three cells: a SQL query, a Python model, a publish call. The results become governed attributes with survivorship and review.
3
kinds of cell: SQL, Python, notes
product
0
export steps to analyse master data with transactions
product
100%
of workbench reads under the same policy and audit as the grid
product
1
call to publish a result as a governed attribute
product
Step by step

From a question to a published attribute

Open a notebook on a domainThe supplier domain: golden suppliers, purchase orders, receipts, quality incidents, all in the bed.
What runs here

Six jobs analysts do on the bed

Profiling

Ranges, patterns, nulls, candidate keys; proposed column rules.

Match-model work

Label pairs, train, evaluate; features include transaction behaviour, because the transactions are here.

Scoring

Supplier risk, customer churn, item obsolescence, published as attributes.

Forecasting

Demand per family and plant from the bed's history; results as a version in the pivot.

What-if

Move a node, change a weight, re-roll the totals; compare with the live version.

Reconciliation

Source against golden, replica against source, invoiced against booked.

Same policy, same audit

Why analysts and stewards see the same data

A workbench query runs as the analyst, through the same engine as the grid: row filters, column permissions and masks apply, and reads of sensitive columns are logged with a purpose. Publishing a result is a write like any other: rules run, survivorship ranks it, the audit trail records the model version behind it.

Models run inside the estate; row data stays where it is. A frontier model can be called for structure and metadata under the boundary described on the AI-assisted page.

Identitythe analyst's own; grants per domain
Readspolicy applied; sensitive reads logged
Publishrules · survivorship · audit · review
Computein the estate; SQL pushed down; Python on results
Versionsnotebooks, models and evaluation sets versioned
Terms on this page

WorkbenchThe place for SQL, Python and statistics on the bed under the same security, with results publishable as attributes.
ProfilingMeasuring a table's values: ranges, patterns, nulls, candidate keys.
Published attributeA value written by the workbench to a golden record, with a rule id, survivorship and review.
Questions

Which languages and libraries?

SQL and Python with the common data-science libraries; R through the same runner on request.

Can a notebook be scheduled?

Yes, as a job with a version, an owner and a result check, with the same audit as a rule run.

Does the workbench replace the lakehouse notebooks?

It sits beside them. Deep engineering stays in your lakehouse; the workbench is for analysis that must run under master-data governance and publish back.

See your first domain built on your data

Bring two systems that hold the same customers, suppliers or materials. In one working day we connect them, run the match on live rows, review the queue with your steward and publish the first golden view with its API.