Introducing Studio
Studio is the workspace area for uploading datasets, creating no-code modelling experiments, and reviewing results.
Use Studio when you want to explore time-series data, test modelling approaches, compare experiment outputs, and generate insights without building a full automated Pipeline workflow.
Studio workflow
Studio is designed for interactive experimentation. It helps you move from uploaded time-series data to modelling results that can be reviewed, compared, and refined before anything is used operationally.
What Studio does
Studio supports the interactive modelling workflow:
-
Uploads datasets - Add CSV datasets containing time-series data and review their structure before modelling.
-
Validates data - Check uploaded datasets for timestamp formats, missing values, row limits, and other issues that may affect modelling.
-
Creates experiments - Choose an experiment type, select a target, configure predictors, set the forecast horizon, and adjust model settings.
-
Runs models - Build models for forecasting, anomaly detection, causal links, or soft sensor workflows.
-
Shows results - Review outputs such as backtesting results, predictions, metrics, feature importance, predictor importance, anomalies, or inferred values depending on the experiment type.
Sections
Studio is organised into the following sections:
- Datasets - Upload, validate, review, and manage datasets used for Studio experiments.
- Experiments - Create modelling experiments, configure targets and predictors, adjust settings, and iterate on previous experiments.
- Results - Review the outputs generated by completed experiments, including model results, metrics, and importance plots.