How many times have you been in a position where you’d like to, for example, have the Map widget or the Devices Table to only display devices that meet a certain condition, either based on its variable’s values or context? Or have the Metric-type widgets (Bar chart, Battery, Gauge, Indicator, Metric, Tank, Temperature, and Variables Table) show their last or aggregated values after filtering them based on a value or context condition?

Well, with Dashboards filters, applying such conditions to your devices’ data is now possible.

There are 2 types of filters, value and context. The first type filters the data displayed in the dashboard based on values, whereas the second uses variable context as the filter input.

Let’s see an example of both of them. To do this, take the following raw time-series data as an example:

First, if we apply a value filter that brings back values greater or equal to 50 (≥ 50), the resulted time series is as shown below. Note that, in the time span, only the values greater or equal to 50 are retrieved.

If we now apply a context filter to the initial time series, so it only retrieves values whose “status” context key is equal to “too-hot”, the resulting time series will be:

The above examples only retrieve the raw time series after applying the filter, but Ubidots’ widgets also support aggregations such as average, max, min, sum, or count. For such cases, the aggregation will be computed after applying the filter to the raw time-series data. For example, let’s assume an “average” aggregation for a value-based filter condition greater or equal to 50 (≥ 50). The resulting time series will be:
​

And the average will be simply 75.

To create filters in your dashboards, simply head to the Dashboard’s settings, and you’ll find the option to create either of the types of filters down at the bottom of the Settings tab:

The below steps describe how to set up a value-type filter:
​
​Step 1: Select the Value option.
​Step 2: Fill the 3 options as follows:

Step 3: Click on Save.
​
Here’s how a value filter configuration that takes as input a variable labeled as “temperature”, looks like:

From the dashboard, you’ll see a new dropdown option name “Temperature” with a half-filled thermometer icon, that after clicking it, it allows entering a filter with the options:

The below steps describe how to set up a context-type filter:
​
​Step 1: Select the Context option.
​Step 2: Fill the 3 options as follows:

Step 3: Click on Save.
​
Here’s how a context filter configuration that takes as input a variable labeled as “temperature” and a its “status” context key, looks like:

From the dashboard, you’ll see a new dropdown option name “Status” with a thermometer icon, that after clicking it, it allows selecting among the different context value options:

Metric-type widgets are:

The impact of filters on these widgets will depend on their “aggregation” configuration, as follow:

​– Last value: will retrieve the first value, from newest to oldest, to match the filter condition in the selected time span.
​– Aggregation: if average, min, max, sum or count is selected, the computation will be applied to the output of the filter in the selected time span.
​

The impact of filters on the Device table widget will depend on the configuration of the columns:
​
​Context data column: If a context filter is applied, the device table widget will retrieve the first value of the context key associated, from newest to oldest which matches the filter context condition in the selected time span.
​
​Value column: The impact of filters will depend on their “aggregation” configuration, as follow:

​– Last value: will retrieve the first value, from newest to oldest, to match the filter condition in the selected time span.
​– Aggregation: if average, min, max, sum or count is selected, the computation will be applied to the output of the filter in the selected time span.

The widgets can be applied to widgets that visualize time series data, such widgets can be:

The filters on these widgets will show the line chart or histogram of the resulting time series after the filters have been applied.