Comment on page
Perform powerful calculations across multiple data sets
The easiest way to perform aggregations on one or multiple data sets is by using the Metric Builder in the Analysis window of Marple.
Marple Analysis Window
The Metric builder consists of 3 components:
- Group by
In the metric part of the builder, you need to define the aggregation method and signal you want to apply it to. You can define multiple metrics.
Currently, the following aggregation methods are available:
In the filter section, you can define the conditions on which you want the metric to be calculated. Setting a filter is optional. If you define multiple filters, they all need to be true (eg. filter1 AND filter2).
In the group by section, you can define how the data needs to be segmented. There are three options:
- All data sets
- By dataset
- By signal
Merge all data and perform the calculations as if it was 1 data set. 1 result per metric will be returned.
Perform the calculations per dataset individually. For each data set, the results of the metrics will be returned.
This is where things become a bit more complex. You can also select the results to be grouped based on a signal. Select a signal and how many buckets you want across a certain range. If you have multiple data sets loaded, the results of all data sets will be merged.
Group by signal
PS: Be careful if you are dealing with integers, you might need to move around the boundaries in order to get the result you want.
The result can be displayed in 3 different ways:
- Bar Chart
- Trend Line
The raw results of the calculation defined in the metric builder can be shown in a table format. In this case, we grouped on a signal, so the first column shows the bucket range. Note that you can sort the table on every column by clicking on the column title.
Result in table format
Often the results can be visualised in a bar chart. Note that you need to select the correct metric that you want to see in the bar chart.
Result in bar chart
The trend line shows how a certain metric evolves over time. This option is only available when you group by dataset and when you use absolute timestamps in your data.
Trend line of a metric
In order to get data in the metric builder, you need to add data from the Library and press 'Analyse' instead of 'Visualise'. Even if you have data open in the Visualisation view, you still need to add them to the Analysis view as well.