Analytics and Outlier Filtering
Environmental sensing and high resolution monitoring can sometimes generate data points called outliers. Outliers are data points that are very distant from other observed values (for example if Temperature is consistently 21C and then for 1 reading the temperature rose to 10000C). Typically, outliers are due to system error and do not represent actual observed values. Error can stem from factors outside of your control (such as electrical noise, a buffer overflow, radio interference, etc). Outliers can cause significant problems with statistical analysis, therefore it is sometimes useful to filter outliers.
A visual example of outlier filtering is shown in the image below. The red line is the unfiltered data with observation errors. The orange line is the filtered dataset (offset for legibility).
You can enable outlier filtering in your analytics queries. Experiment with different IQR multiples to get the best results that suit your needs.
For more information on how to enabled Outlier filters in your Analytics queries, click here: https://www2.aretas.ca/knowledge-base/outlier-filtering/