Uniqueness analysis over time
Learn about the data in the Uniqueness analysis chart in the DQ dashboard.
Monitor the quality of data in your Reltio tenant on the Real-Time Data Quality (DQ) dashboard. After selecting an entity and attributes on the DQ dashboard, you can find out about unique and common values of attributes on this Uniqueness analysis over time DQ chart. For more details, see topics Real-Time Data Quality dashboard charts and Analyze DQ dashboard charts.
You can also view anomalies in your data. Anomalies are deviations from the expected pattern in the data. In this chart, you can view data points that are recorded at time intervals. These intervals generally vary but you can recognize changes in the data and identify repeating patterns. Data points not conforming to established patterns are flagged as potential anomalies.
If you're a Data Analyst, Data Steward, or Business User, this information is for you! Use the metrics in this chart to identify unique profiles where you may need to define specific attributes in profiles.
Data definition
| Data label | Data definition |
|---|---|
| Uniqueness analysis over time | The relative number and percentage (up to 3 decimal places) of unique attributes in the entity that contains the values selected in the Attribute panel:
|
| Time period | Indicates the time frame for the attribute data:
Note: If the date range changes and the current date is no longer in range, a date range displays instead of Today. For details on specifying a time period, see topic Analyze DQ dashboard charts. |
| % of unique distinct values | Identifies the percentage (up to 3 decimal places) of unique attributes for the selected Attribute. |
| % changed over time | Identifies the percentage (up to 3 decimal places) of unique attributes changed over the selected time period for the selected Attribute. |
No data
You can view at a glance, by hovering over the graph, when data is not available for a particular period.
Anomaly detection
You can view error bands in this chart to determine anomalies in your data. For each of these data points, you can view an error band. An error band provides the upper and lower limit for the trend observed in the data. If the actual values are above or below the limits, it indicates an anomaly or deviation from the expected value. This information helps you resolve issues in your data.
How does this work? Well, the anomaly model uses historical data for a time period to make forecasts or predictions. For example, assume you opted to view data for the last seven days. The data from the start date of the selected time period to the day before the current date is considered to make the forecast.
The actual data is then compared to the forecasts. If the actual data is within the error band or limit, it indicates normal behavior. If the actual data is above or below the limits, it is considered an anomaly.
Select the i next to the chart name to view details about the chart, as shown below:
In the chart, you can hover over any data point to view the anomaly details . A sample is given below: