It is possible to calculate the average per period of a given indicator over time in the *Explorer* graph. In the example below, we obtain the average number of visits per hour over a day.

This calculation can be adjusted by changing the analysis period, the granularity of the curve (minute, hour, day, week, month) and the metric analysed.

To obtain this information in a table or as a KPI, it is necessary to create a custom metric.

For example, to calculate the average number of visits per week during a varying analysis period, the recommended calculation is as follows:

*(Total number of visits during the period/Number of days with visits during the period)*7*

The calculation of an average being based on the analysis period, its value may vary when the analysis period changes. It is recommended that long analysis periods are chosen to obtain a reliable average.