How exactly does the analysis for discrete stories work? This article will explain the different analytic types that power each sentence in your narrative.

Discrete stories allow the user to compare values and understand the distribution of the data. The narrative will analyze distribution, averages, totals, and groupings/clusters across the data.

Let's look at the sample narrative below, which measures Sales by State. This is a one dimension, one measures story.

We can see the narrative talks about total sales, the different states driving those sales, the states with the highest and lowest sales, as well as the distribution and concentration of sales. These ideas are all powered by different analytic types that help make the narrative interesting and actionable.

The first bullet will always call out the total value of your measures. In this case, it is total sales across all the states.

The second bullet calls out dimension drivers:

• "The Sales of \$2.3 million was driven by California with \$457,731, New York with \$310,914 and Texas with \$170,187."

Drivers are essentially the measures that contributed most to the total value. You can edit the number of drivers you want to include in your story.

The third bullet uses range analysis to call out the smallest and largest values, as well as the difference and average:

• "The minimum value is \$920 (North Dakota) and the maximum is \$457,731 (California), a difference of \$456,811, averaging \$46,855."

The next two bullets analyze the distribution of your data. This will analyze how the data is skewed, averages, medians, and concentration of data (if any). This answers questions like: how balanced are these grouped variables compared to one another?

• "The distribution is positively skewed as the average of \$46,885 is greater than the median of \$22,207."
• "Sales is somewhat concentrated with 13 of the 49 states (27%) representing 76% of the total."

The next bullet in this story uses clustering analytics to call out any measures that can be grouped together. Clustering answers questions like: are there any distinct groups that stick out in the data?

• "The top two states account for over a quarter (33%) of overall Sales."

The last bullet mentions any notable outliers:

• "New York stood out with a high Sales value."

As you can see, discrete story analytics includes distribution, averages, totals, and groupings/clusters.

Note the number of bullets in your narrative can vary depending on your chosen verbosity. The higher the verbosity, the more sentences will be written using the analytics types outlined above.