The data every model downstream depends on

A statistics office is the upstream of the whole estate. If the national accounts, the SAM and the household survey are not governed, versioned and published properly, every model in every ministry inherits the problem silently. EDIS makes that dependency explicit and manageable.

Request a demonstration Build your package

The workflow

From collection to the calibrated model

Data is versioned, quality-controlled and published once, with provenance attached, and every downstream model records which vintage it used.

Model chainStatistics

01Collection and administrative dataSurveys, censuses and administrative registers, with coverage and quality recorded

02Quality controlValidation, outlier detection and the documented treatment of missing data

03Metadata and provenanceSource, vintage, method and known limitation carried with every series

04National accountsThe accounts compiled, reconciled and published as a governed dataset

05Household survey analysisWeights, strata and clusters handled properly, with design-based errors

06SAM and calibrationThe social accounting matrix and the parameters the CGE estate calibrates on

07Data APIsOne governed publication point for ministries, the central bank and researchers

08Downstream modelsEvery model records the data vintage it ran on

Questions it answers

The questions a statistics office brings

Governance

Who is using which vintage of our data?

Every model run records the dataset and version it used, so a revision can be traced to everything it affects.

Quality

Where is our data too thin to support the model?

Quality control is explicit, and EDIS states where the data cannot support a model rather than substituting a proxy.

Publication

How do we stop serving spreadsheets by email?

Governed data APIs give ministries and researchers one publication point, within permissions.

Calibration

Can our SAM support the CGE the ministry wants to run?

The calibration tools show what the SAM supports and what it does not, before a model is built on it.

Surveys

Are the poverty numbers computed correctly?

Household analysis handles weights, strata and clusters with design-based standard errors, and states the accounting unit and the denominator.

Revision

What happens when we revise the accounts?

The revision propagates as a new vintage, and every downstream result says which vintage it used.

What is included

What a statistics deployment carries

The economic data platform and database architecture
National accounts compilation and reconciliation
Household survey analysis with design-based inference
Administrative data integration
Data quality, metadata and provenance management
SAM construction and CDCGE calibration tools
Governed data APIs for the rest of government
GIS and spatial data infrastructure
EDIS statistical agents
Institutional workspaces, roles and an audit trail
Technical training and certification for statistical staff
An annual review, and priority support with a named team

Request a formal proposal Compare entitlements

Recommended licence

From per year

Institutional pricing is contracted and reflects model configuration, national data integration, calibration, training and support rather than a seat count. Implementation is quoted separately, and development access pricing is applied where the country’s income classification allows.

Next step

Start with the dataset the ministries argue about

We show how it would be governed, versioned and published, and what changes downstream when it is.

Request a demonstration Request a sandbox demo All government solutions