Effect of "Terms" aggregation when it has been selected as the aggregation for splitting bars

Elastic Stack Kibana
1

Hi

In the first bar chart I have shown average of sales per day during a period and in the second one I have used "Split bars" as sub_bucket in order to show average of sales per day per type of customers during the period. But the result has been completely changed. The range of y.axis in the first graph is between 0 , 1000 . But in the second one is between 0 , 5500. For some of days , this amount extremely has been changed.

1.1.jpg1148×769 35 KB

1.2.jpg1145×776 31.5 KB

I'm agreed that it's logically some of bars will see an increase in its value because the average will be calculated in other way (For example : There is a type of customer which has very low number of custmers ( such as 5) but each of them has high value in their sale.) .

The point is , as you can see the peak of first graph contains those customers which "only" belong to "one" group (type) of customers. Even in this case , a huge increase can be seen there.

Is there any one can help me with this ?

2

Hi Saba,

Could you please post a screenshot of the aggregations used in the charts?

Could you also try going into the options for the chart and changed the 'Bar Mode' from 'stacked' to 'grouped' and see what that looks like. Stacking the averages probably gives a misleading view.

Another idea is to use total sales instead of average sales. The totals should stay the same after you split them by customer type.

Thanks,
Lee