AI agent - government data

EHOBA walkthrough

A scrubbed walkthrough of a desktop agent for the monthly hotel occupancy survey. It shows the flow from synthetic CSV review to processing, correction review, and generated outputs.

Privacy status: this page uses a sandbox with fictional hotel names, fictional 99-series identifiers, and no real survey values. The operational app and real government datasets are not deployed publicly.
Animated scrubbed EHOBA walkthrough showing synthetic input review, processing, corrections, and generated outputs
1

Review input

A synthetic LimeSurvey-shaped export is checked against a fictional establishment frame before processing.

2

Run rules

The processor applies fixed validation rules and writes a repeatable run record.

3

Check C4

Correction cases are displayed for review without exposing raw survey values.

4

Write outputs

The sandbox run creates PDFs, spreadsheets, text summaries, and history files.

Synthetic EHOBA input review screen
Input review with fictional hotel names and 99-series identifiers.
Synthetic EHOBA processing run screen
Processing run from the sandbox: 4 synthetic establishments processed, 2 C4.
Synthetic EHOBA correction review screen
Correction review shows rule categories, not raw government survey values.
Synthetic EHOBA generated outputs screen
Generated outputs are represented with fictional file names and identifiers only.