Trend Binary Sensor

The trend platform allows you to create sensors which show the trend of numeric state orstate_attributes from other entities. This sensor requires at least two updates of the underlying sensor to establish a trend. Thus it can take some time to show an accurate state. It can be useful as part of automations, where you want to base an action on a trend.

To enable Trend binary sensors in your installation, add the following to your configuration.yaml file:

# Example configuration.yaml entry
  - platform: trend
        entity_id: sensor.cpu_speed

Configuration variables:

  • sensors array (Required): List of your sensors.
    • entity_id (Required): The entity that this sensor tracks.
    • attribute (Optional): The attribute of the entity that this sensor tracks. If no attribute is specified then the sensor will track the state.
    • device_class (Optional): The type/class of the sensor to set the icon in the frontend.
    • friendly_name (Optional): Name to use in the Frontend.
    • invert (Optional): Invert the result. A true value would mean descending rather than ascending. Defaults to False
    • max_samples (Optional): Limit the maximum number of stored samples. Defaults to 2.
    • min_gradient (Optional): The minimum rate at which the observed value must be changing for this sensor to switch on. The gradient is measured in sensor units per second. Defaults to 0.0
    • sample_duration (Optional): The duration in seconds to store samples for. Samples older than this value will be discarded. Defaults to 0

Using Multiple Samples

If the optional sample_duration and max_samples parameters are specified then multiple samples can be stored and used to detect long-term trends.

Each time the state changes, a new sample is stored along with the sample time. Samples older than sample_duration seconds will be discarded.

A trend line is then fitted to the available samples, and the gradient of this line is compared to min_gradient to determine the state of the trend sensor. The gradient is measured in sensor units per second - so if you want to know when the temperature is falling by 2 degrees per hour, use a gradient of (-2) / (60 x 60) = -0.00055

The current number of stored samples is displayed on the States page.


In this section you find some real life examples of how to use this sensor.

This example indicates true if the sun is still rising:

  - platform: trend
        entity_id: sun.sun
        attribute: elevation

This example creates two sensors to indicate whether the temperature is rising or falling at a rate of at least 3 degrees an hour, and collects samples over a two hour period:

  - platform: trend
        entity_id: sensor.outside_temperature
        sample_duration: 7200
        min_gradient: -0.0008
        device_class: cold

        entity_id: sensor.outside_temperature
        sample_duration: 7200
        min_gradient: 0.0008
        device_class: heat