Data validation
For projects that have a tracking plan configured and already have reported data, you can validate whether the quality of the stored data meets expectations. During data validation, the system batch-scans the data quality within the specified scope and identifies events or properties that don't meet the validation criteria.
Validation mainly covers:
- Differences between the reported data and the data processing rules defined in the tracking plan.
- Null rates of reported properties. A property is identified as abnormal when its null rate is greater than the specified threshold.
Note: Data validation covers events and custom properties. Preset properties are not validated.
Configure data validation rules
Click Run validation test in the upper-right corner of the page to configure validation rules. You can choose to validate event data or user data.
(1) Validate reported event data
You can configure the scope of this validation by time range, event range, and filter conditions:
- Time Range: The time range of the events to validate. It defaults to yesterday and can be customized. Note: This time refers to the #time field in the reported data and is not offset by the project time zone.
- Select events: The events to validate. It defaults to all events, and you can customize the event range, for example, select the tracking events added in a new version for validation.
- Filter: You can filter event data. If you collect and report data through multiple client SDKs, you can validate only the data from a specific source.
Set the property null rate validation rules, including the null rate threshold and the counting rule:
- Null rate threshold: When the null rate of a reported property exceeds the threshold, it is shown as abnormal in the validation results.
- Counting rule: You can choose whether to count empty strings into the property null rate.
(2) Validate user data
Set the user data scope, property null rate validation rules, and other settings for validation. The steps are the same as for validating event data.
2. Validation results
After the validation calculation is complete, you can view the validation report.
The validation report includes information such as Content, Condition, Status, and Conducted by.
The validation details page also shows the validation results of the data
Validation result statuses for Events:
| Validation results | Description |
|---|---|
| Success | The reported data matches the tracking plan, and all property null rates are lower than the set threshold |
| No data | No event data is reported within the selected validation scope |
| Abnormal | An event is marked as abnormal in any of the following cases: 1. Events that are not in the tracking plan exist 2. Properties that are not in the tracking plan but are actually reported exist 3. Properties are missing, that is, properties that are in the tracking plan but are not actually reported 4. The type of a reported property is inconsistent with the property type in the tracking plan 5. The property null rate exceeds the set threshold |
Validation results for Users:
| Validation results | Description |
|---|---|
| Success | The reported property data matches the tracking plan, and property null rates are lower than the set threshold |
| Abnormal | A user property is marked as abnormal in any of the following cases: 1. User properties that are not in the tracking plan 2. Missing properties: user properties that are in the tracking plan but are not actually reported 3. The type of a reported property is inconsistent with the property type in the tracking plan 4. The property null rate exceeds the set threshold |