What you can do in Predict
Data
- Upload CSV time-series files (date + numeric target).
- Browse Data Sources: filename, rows, columns, workspace (
analyze_csv/predict), created date. - Search and page through uploads.
Forecasting
- Create prediction jobs (train + forecast) with Prophet, ARIMA, or XGBoost.
- Choose daily, monthly, or hourly horizons and period count.
- Track job status: PENDING → PROGRESS → SUCCESS / FAILURE.
- View MSE and RMSE (error metrics — lower is better).
- Open job results on the Predictions tab and export when enabled.
Project & account
- Studio project selector (example: AI Project).
- Dark/light theme.
- You see jobs and datasets for your account / project context.
Reporting & control
- Dashboard: job counts, users, latest error metrics, recent jobs.
- Audit Logs: timestamp, actor, action, target, IP; search and export.
Optional: API
Same capabilities exist as REST (Swagger on the Predict API host). See API if you are not using the Studio UI: analyze_csv, predict → job_id, jobs status/result, save_model, predict_with_loaded_model, datasets.