Skip to content
English
Contact us
Contact Us
  • There are no suggestions because the search field is empty.

Understand Data Completeness

Learn what the Data Completeness metric measures, how it is calculated, and how to interpret it in Analytics.

Table of Contents

What Data Completeness Measures

Data completeness tells you how much of a sensor's expected data actually arrived over a selected time period. It is shown as a percentage: the number of readings received divided by the number of readings expected, based on how often the sensor is configured to report.

A high percentage means the sensor reported consistently. A low percentage means readings are missing, which can happen when a sensor loses connectivity, runs low on battery, or is out of gateway range.

Note

Data completeness is available for active sensors that report environmental measurements on a regular schedule, such as temperature, humidity, light, and UV. It is not calculated for event-based sensors such as leak detectors (which only send messages when they are wet).

How Completeness Is Calculated

Completeness compares two numbers over the time period you select:

  • Expected readings — how many readings the sensor should have reported, based on its reporting interval. For example, a sensor reporting every 15 minutes is expected to send 96 readings per day.
  • Received readings — how many readings are actually stored. This includes readings captured by the bluetooth gap fill feature.

The result is the received count divided by the expected count, shown as a percentage from 0 to 100.

Completeness is calculated for a single sensor over a time period of up to one year.

Known limitations

When interpreting the data completeness metric, bear the following limitations in mind:

  • The KPI is not aware of when the sensor first started transmitting readings. Do not use the KPI for time periods that start before the sensor was installed, as this will give a result that is worse than reality (for example, if the sensor was installed yesterday, don't use the KPI with a time period of "last week").
  • The sensor calculates expected readings based on a single reporting interval. If you have changed the reporting interval for your sensor, the data completeness metric will be incorrect for periods with a different reporting interval.
  • Data completeness is not calculated for all sensor types - for a list of exceptions, see the "When Completeness Cannot be Calculated" section below.

 

Read the Completeness Rating

Data completeness appears in Analytics when you view a single sensor.

  1. Open the CUSTOMIZE menu

  2. Turn on Show Data Completeness to display the completeness gauge alongside the sensor's readings.

The completeness percentage is shown as a gauge with a plain language rating so you can assess data quality at a glance:

Rating Completeness

Excellent

90% or higher

Very Good

80% to 89%

Good

70% to 79%

Fair

50% to 69%

Poor

Below 50%

Tip

80% or higher is very good. At that level, you have enough data to trust trends, averages, and alerts for the period. Ratings below that suggest gaps that may affect your analysis. 100% is not fully necessary for most general purposes and may not always be achievable for sensors that have weak signal strength or are in noisy environments.

When Completeness Cannot Be Calculated

In some cases, a completeness percentage is not shown. This happens when:

  • The sensor is inactive.
  • The sensor does not report continuous environmental measurements.
  • The sensor's expected reporting interval cannot be determined; this currently includes imported sensors.
  • The selected time period is longer than one year.

In these cases, Conserv Cloud does not display a completeness value rather than show a number that could be misleading.

Merged Sensors

If a sensor has replaced an older, merged sensor, completeness includes the full combined history so it matches the readings shown on the chart.

Caution

As described above, if a merged sensor previously reported on a different interval, the completeness value for periods that span the merge will not be precise. Conserv Cloud flags this on the completeness gauge.