Workflow
A sequence of activities: it acquires data, checks it, performs operations and preserves the result.
Example: every morning it reads weather data and prepares an updated card.
Three ideas in plain language
What the system does, without jargon.
Automation and RPA
Software repeats calculations or digital activities — downloading files, comparing values, updating registers — without manual repetition.
Forecasting and classification
A model may estimate a value or probability, or assign a case to a category. The result supports judgement but does not decide on its own.
A complete example
From source to result, step by step.
- 1 · CollectionThe system acquires values from public sources.
- 2 · ChecksIt verifies freshness, format, missing values and obvious anomalies.
- 3 · CalculationIt combines data and calculates changes or indicators.
- 4 · Model, where usefulIt produces an estimate or classification without presenting it as certainty.
- 5 · Final gateIf data are insufficient, the workflow stops or requests review.
- 6 · Output and traceIt prepares a chart, map or report and preserves sources, date, version and errors.
Examples developed in the laboratory
Three different uses of data, automation and AI.
The homepage shows a selection. The full portfolio documents sources, controls, results and limitations for every workflow.
WF-9 · World economy · AI/ML · research
World Economy Engine
Integrates macro and market data into a Global Economy Pulse, time-causal regimes, a structural graph, scenarios and forecasts with prequential selection, benchmarks and uncertainty intervals.
WF-10 · Geoscience · monitoring · RPA/ML
Etna Sentinel
Automated relay of official VONA notices plus an educational wind-transport simulation combining INGV, Open-Meteo, NASA FIRMS/GIBS, pressure-level trajectories, infrastructure proximity and safety gates.
WF-7 · Geoscience · AI/ML
Etna Forecast
24-hour and 5-day probabilistic nowcasting with public multimodal data, strong baselines, a modality gate and declared degradation.
Guided path
Choose how far to explore.
Start with essential terms and continue, when useful, towards method, architecture and validation.
Why verifiable
The result is not presented on its own.
Repeatable
Activities are defined and can be executed through the same sequence.
Inspectable
Sources, checks, version and limitations remain visible.
Prudent
Insufficient data, uncertainty and known failure conditions are not hidden.
Describe a non-confidential question
State the expected result and context of use without sending personal data, credentials, internal documents or confidential information.