This article will explain how to configure and update variance thresholds for Rio’s automated anomaly and variance detection
What are anomalies and variances?
| Category | Checks | Rule |
| Anomaly Detection | Meters or locations with zero values | On upload |
| Anomaly Detection | Data received outside of location end dates | On upload |
| Variance Detection |
Variances compared to the same month, prior year For example: % change between June 2023 and June 2022 |
On upload |
| Variance Detection |
Variances compared to the prior month of the same year For example: % change between June 2023 and May 2023 |
On upload |
| Anomaly Detection | Anomalies based on overlapping transactions within the same time period | On upload |
As part of Rio’s automated anomaly detection, Rio will scan your data upon upload for the issues outlined in the table above. These checks are grouped as either a variance or anomaly.
Anomaly: Overlapping transactions or data received after occupation end dates, these flags allow you to review for duplicates or make any changes in configuration which most accurately reflect the activity of your operations.
Variance: Calculated as a percentage change compared to historic data, a variance flags transactions which vary significantly from historic data trends.
For example, This means for every electricity upload, Rio will check for:
- Annual Variances: Calculate the percentage change compared to the same month of the previous year (12 month comparison)
- Ex. You just uploaded your Jan 2023 Data. An annual variance is Jan 2022 compared to Jan 2023 (Rio is calculating: How much higher or lower is Jan 23 compared to Jan 22?)
and
- Monthly Variances: Calculate the percentage change compared to the previous month, same year (1 month comparison).
- Ex. You just uploaded your Jan 2023 Data. A monthly variance is Jan 2023 compared to Dec 2022 (Rio is calculating: How much higher or lower is Jan 23 compared to Dec 22?)
If a transaction is above the configured positive variance threshold (analysing for increases in data) or below the negative variance threshold (analysing for decreases in data) Rio will flag this for review in the Data section of the platform.
Note: Rio only flags these variances. This will not prevent high or low data from entering the platform
Configuring Variance Thresholds
Anomaly detection requires no configuration, as these checks are performed automatically once your data tracker is turned on. However, variances are customisable and will require a bit of configuration to set up.
So how does Rio know to flag a variance as too high or too low? That’s up to you and your organisation to determine!
During onboarding, your Rio Consultant will ask if you would like to configure your variances. These can be configured at the utility level for your organisation.
Only Business Admins and Rio Champion user types will be able to configure and edit variances:
To configure or edit a variance threshold navigate to configuration>data tracker configuration>anomalies thresholds
Then select “+Add Data Tracker Configuration” to create a new threshold
Indicate the threshold percentage and utility this applies to, then select “create”
All thresholds will appear here in configuration.
If you would like to change a threshold, select the edit/pencil in the far right column, then select update once that figure has been updated.
Variances can also be removed by selecting the bin icon to right of the edit/pencil icon.
Only once a variance is configured will Rio begin analysing your data according to those thresholds.
Viewing and resolving your variances
Now that you've configured your variance thresholds, you're ready to analyse and review your flagged transactions. For more information on how to view and resolved your variances see our support page here: