Docs Data Governance
Data Governance helps discover and repair time-series data quality issues, improving data usability and downstream modeling results. You can select an existing dataset, start governance, compare results before and after governance, and review historical records.
Start Governance
On the Data Governance page, select the dataset to process and click Start Governance.
After governance completes, the page shows the latest governance time and a before/after comparison, including timestamp accuracy, missing value rate, stable series ratio, anomaly rate, and overall score. Each metric compares the result before and after governance and indicates whether it increased, decreased, or remained unchanged, so you can quickly judge the effect.
Governance Effect
The comparison area shows the latest governance result:
| Metric | Description |
|---|---|
| Timestamp accuracy | Correctness of timestamp format and ordering |
| Missing value rate | Ratio of missing points in the data |
| Stable series ratio | Ratio of series segments that are stable and suitable for modeling |
| Anomaly rate | Ratio of outliers or abnormal fluctuations |
| Overall score | Overall data quality score after governance |
Interpret metric changes with business context. Some metrics have different preferred directions: missing value rate and anomaly rate are usually better when lower, while timestamp accuracy is usually better when higher.
Governance Records
Governance records keep historical tasks. You can search by dataset name and refresh the list to get the latest results.
The list includes:
| Field | Description |
|---|---|
| Record number | Identifier of the governance record |
| Dataset name | Original dataset name |
| Governed dataset name | Dataset generated after governance |
| Overall score | Quality score after governance; if the post-governance evaluation cannot be completed, the record shows Post-governance evaluation failed |
| Created at | Time when the governance task was created |
| Duration | Execution time of the governance task |
| Actions | View governance details or delete the record |
View Governance Details
Click View for a governance record to open the detail page. The page includes the governance process, task information, improvement, and detailed results.
The Governance Process area shows the current task status and overall progress. When governance is complete, the status is Completed and progress is 100%.
The task overview shows the governance task ID, dataset name, governance time, duration, score before governance, and score after governance. It also highlights the score change, calculated as the score after governance minus the score before governance. The result can be positive, zero, or negative. A lower score does not mean that the governance task failed; interpret it together with the individual metrics and business context.
The Details table shows the following information for each object:
| Field | Description |
|---|---|
| Object name | Name of the object in the dataset |
| Series count | Number of time series contained in the object |
| Total points | Total number of data points contained in the object |
| Score before governance | Data quality score before governance |
| Score after governance | Data quality score after governance |
| Improvement | Score after governance minus score before governance; the result can be positive, zero, or negative |
Click the expand control to the left of an object to inspect its series names and the before score, after score, and improvement for each series. For further analysis, export the results as CSV or switch the table to full-screen view.
Relationship with Datasets
In the dataset list, each dataset can be used for both Data Evaluation and Data Governance. After governance completes, the system generates a governed dataset that can be used for data evaluation, model training, or forecasting analysis.