1.1 Discrete - Single Dimension, Single Measure


Introduction 

 

The content in this section makes up the entire narrative when there is a single dimension and a single measure.

 

Note that this content also makes up the drill-down portions of the multi-dimension, single measure. In other words, each drill-down in the more complex narrative, can be thought of independently as a simple single dimension, single measure narrative.

 

Please reference - Discrete Analytic Buckets

  • Distribution - Describes how values are distributed among the various dimension entities. Includes normal distribution, skewness, and concentration. Found in discrete story types only.
  • Clustering - Describes notable groupings based on the values of the dimension entities. Found in discrete story types only.
  • Correlation - Compares two measures to identify any relevant relationships (positive or negative) between the two variables. Found in discrete and continuous story types.
  • General - This refers to any narrative content not associated with any particular bucket listed above. This bucket includes headlines, intros, descriptive content, or content that doesn't require complex mathematical/statistical models. Found in all story types.

 

 

Narrative Content  

 

Headline - states the measures and dimensions written about in the narrative

e.g. - This chart measures Sales Revenue by Product Category. 

  • - Analytics Bucket: General
  • - data-section-id="intro"

 

Measure Aggregation - Sums or averages the measure value across all dimension entities. Total (sum) vs average is dependent on the measure characterizations that measure, specifically, “do you want to know the cumulative total?”

e.g. Total Sales Revenue is $1.01 billion across all eight product categories. 

  • - Analytics Bucket: General
  • - data-content-id="measure-totals"

 

Range - presents the dimension entities with the lowest and highest values. It also presents the difference (range) and the average. Note, it will not present the average if measure characterization has been set to average because the average will have already been called out in the Measure Aggregation sentence.

e.g. - The distribution ranges from $250,147 (MISC. NON-INVENTORY) to $360.8 million (HOME ELECTRONICS), a difference of $360.6 million, averaging $126.3 million. 

  • - Analytics Bucket: General
  • - data-content-id="range"

 

Mean and Median - presents the means and the median. Note, it will not present the average/mean if measure characterization has been set to average, because the average will have already been called out in the Measure Aggregation sentence.

e.g. - The average Sales Revenue per product category is $126.3 million and the median is $85.8 million. 

  • - Analytics Bucket: General
  • - data-content-id="mean-median"

 

Mode - presents the mode, or the most commonly occurring value(s).

e.g. - The most common values are $250,147, $10.6 million and $57.4 million which each occurred one time. 

  • - Analytics Bucket: General
  • - data-content-id="mode"

 

Concentration - makes an assessment of how concentrated the distribution is among the top dimension entities. This analysis compares the proportion of dimension values that have a measure value above the mean (e.g. 38% of the entities, or 3 out of 8) to total value that those entities represent (e.g. 76%, the total revenue of those top three entities as a percentage of total revenue across all eight entities). Word choice in describing the level of concentration changes based on the strength of that ratio. There is also a follow-up sentence/bullet explaining the smaller entities (i.e. the remaining five entities)

e.g. - Sales Revenue is somewhat concentrated with three of the eight product categories (38%) representing 76% of the total. On the other hand, the five smallest product categories (63% of the total) represent 24% of the total Sales Revenue. 

  • - Analytics Bucket: Distribution
  • - data-content-id="concentration"

 

Distribution - calls out whether or not the data follows a Normal Distribution at a 90% confidence level. Is there is a normal distribution, it presents the mean and standard deviation. If relevant, it will provide an example of the proportion of the dimensions entities expected to fall within n standard deviations of the mean. Alternatively, if there is no normal distribution, it calls out the positive or negative skew and comments on the average and median in the context of skewness. 

e.g. - The distribution is positively skewed as the average of $10.7 million is much greater than the median of $3.2 million.

  • - Analytics Bucket: Distribution
  • - data-content-id="central-tendency"

 

Top Entities - calls out the top dimension entities and presents the proportion of the metric that they represent. If one dimension entity represents more than 50%, it only calls out that one and not the second biggest entity. If the measure characterization is set such that larger values are considered bad, this will do a similar analysis, but focused on the entities with the lowest values.

e.g. - HOME ELECTRONICS accounts for 36% of overall Sales Revenue, and the top two product categories combine for over a half (62%). 

  • - Analytics Bucket: General 
  • - Varies based on measure characterizations
  • - data-content-id="bars-top"

 

Top Entity Comparison - compares the top dimension entity to the next biggest and/or to the average. If the measure characterization is set such that larger values are considered bad, this content will not print.

e.g. - HOME ELECTRONICS ($360.8 million) is almost three times bigger than the average across the eight product categories. General

  • - Analytics Bucket: General
  • - Varies based on measure characterizations
  • - data-content-id="first-over"

 

Outliers - calls out dimension values that lie at least 2 or 3 standard deviations from the mean.

e.g. - Large Screen LCD TVs, Projection TVs, Mattresses, Plasma TVs and Laptops were exceptions with very high Sales Revenue values.

  • - Analytics Bucket: General
  • - data-content-id="measure-outliers"


Was this article helpful?
0 out of 0 found this helpful