Data Quality

Understanding Data Quality from the CORE sensor

When the CORE sensor calculates the core body temperature, it takes a few factors into account to determine the data quality. When the CORE sensor is worn, the internal sensors record the thermal energy transfer values, the skin temperature and (if connected) heart rate signals.

The CORE sensor identifies the 'validity' of many of these signals and sets a data quality score from 1 (poor) to 4 (best) which is used in the CORE App and CORE cloud along with other devices to determine how and if the core body temperature data is displayed.

Factors that impact the data quality are irregularities such as low skin temperature, unusually low or high heart rate or unusual changes in heart rate or if no movement is detected to suggest the CORE is not being worn.

Commonly, lower data quality typically only occurs for short time periods, such as when the CORE sensor is first put on.

How is the data quality is displayed

Before outlining the factors that impact data quality, first we will address how the data quality is represented.

The CORE Cloud web app shows a grey background for a low data quality when the CORE is first worn and later when the skin temperature dips.

The CORE App shows the CORE temp plot in grey for lower data quality at the beginning and end of this workout.

The .CSV file that can be downloaded from the CORE Cloud marks the data quality in the TempQuality.

*Please note that access to the cloud dashboard and the associated CSV download is restricted to purchasers of the CaleraResearch Product and pro account holders*

What factors impact data quality

When any of these factors (conditions) are triggered, the data quality is adjusted.

CORE Sensor is initialising

When the CORE sensor is first worn, it takes a few moments for it to capture the data and start calculating. In these few seconds , the data quality is 0 and the message “Waiting for valid value” is displayed

Accuracy increase over time

When the CORE sensor is first put on, the initial data it receives can be irregular and the algorithm also requires sufficient of data to ‘normalise’ which takes time. By default, when the CORE sensor is first worn, the data quality is set to 1 and then increases over time:

Under 5 minutes: Data quality = 1
5 – 10 minutes: Data quality = 2
10 – 15 minutes: Data quality = 3
Greater than 15 minutes: Data quality = 4

No movement detected

The CORE sensor has an inbuilt accelerometer that detects movement. If no movement is detected, this is one indicator that the CORE sensor may no longer be worn and will impact the data quality score.

After 30 seconds of no movement – Data Quality – 3
After 5 minutes of no movement – Data Quality – 2
After 30 minutes of no movement – Data Quality – 1

Skin Temperature

In addition to recording the thermal energy transfer to calculate the core body temperature, the skin temperature is also recorded. When unexpected temperatures are detected then this impacts the data quality. The skin temperature may still be accurate (for example, extreme cooling or for people with uncommonly low skin temperature) though a low skin temperature can also be an indication that the CORE sensor is not being worn and is, for example, resting on a table and now receiving this surface temperature.

Skin temperature < 28ºC = Data Quality – 1
Skin temperature < 31ºC = Data Quality – 2
Skin temperature < 34ºC = Data Quality – 3
(Higher than 34ºC = Data Quality – 4)

Heart Rate

For sports use and activity with effort, it is recommended to pair the CORE sensor with a heart rate monitor which will ensure a level of accuracy of core body temperature calculations during sport. If the CORE sensor detects a paired heart rate monitor, it uses the data to determine the level of effort to trigger a switch from the default algorithm (suited to everyday activity) to the ‘Sport Mode’ algorithm. In addition, the heart rate also impacts the data quality when the values appear irregular.

The data quality will be set to 1 for a duration of 5 minutes in the following cases:

• If the heart rate detects a jump of over 50bpm in less than 3 seconds

• If the heart rate is higher than 220bpm for more than 15 seconds

• If the heart rate is less than 30bpm for more than 15 seconds

• If the heart rate has no change (exactly the same value) for more than 15 seconds

Interpreting Data Quality

Typically the data quality is an indication how reliable the core body temperature data if the data is suitable for analysis and decision making.

In practice, the CORE sensor should be worn 10-15 minutes prior to starting sport activity.

If the data quality drops during an activity, the skin temperature values can be checked to identify any natural change (e.g. rapidly lowering skin temperature from cooling). In this case, the data quality may be lower as skin temperature is outside the normal/expect range though the core temperature calculation remains valid. A low data quality drop during an activity can also be an important indicator of an issue, such as poor or interrupted heart rate data, which results in poor accuracy of core body temperature.

Some individuals with low skin temperatures or heart rate values may see an impact on the data quality. In some scenarios the data is still valid (even though the data quality is marked lower) while in other scenarios, untypical individual characteristics are affecting the accuracy of the core body temperature calculations, in which case we recommend contacting one of our experts who can review your data.