Quill for Qlik comes with the ability to write insightful narratives even from data that contains missing or null values. This guide will walk through the scenarios that are supported by this feature.
1. Multi-dimension narratives with an aggregated measure and dimension values not present in every period.
In the example below, we have a line chart of revenue by product where several products did not exist for the entire length of the series. So, the yellow 'Dairy' product did not start selling until Jun-06 where the red 'Baking Goods' product did not start until Jan-07.
In this case, the data looks like the below:
You will notice that Jan-06 through May-06 only has two rows each per month. This is because there was no 'Dairy' or 'Baking Goods' categories in those months. The rows just do not exist.
In a previous version of N4Q, the narrative would throw away any incomplete lines. So, 'Dairy' and 'Baking Goods' would not be included in any analysis.
Now, the narrative will look like the below:
Now, the narrative will analyze all four series across the periods where all have overlap. In this case, from Jan-07 through Dec-07.
For each individual series, the narrative will analyze the series in a drill down for the entire length of that particular series. For Dairy, the drill down section will analyze from Jun-06 through Dec-07 while for beverages, it would analyze Jan-06 through Dec-07.
2. Multi-measure narratives with un-aggregated measure values and a single dimension
In the example below, we have a multi-measure line chart with missing values. The blue line only goes from 2004 to 2015 and the red line goes from 2000-2013.
Here, the data is not being aggregated. You can see in the formula for the measure, only the raw value is being taken. We recommend removing all formatting from null value cells (e.g. $) and identifying values as dimensions to prevent aggregation by including brackets (e.g., [direct injuries]).
This results in a tabular data representation like the below:
In this case, like the above, the narrative will compare the series based on the periods of complete overlap and will analyze each individual series across the periods for which it has data.