Automated Insights

This page describes the Automated Insights capability.

Sometimes, a graph alone is not sufficient to convey all the granular meanings hidden under data and digging into it can be very time consuming and requiring specific technical skills.

For this reason, alongside with the Topic Chart, each Topic always comes with a set of additional information, called Automated Insights, that are directly related to it.

They represent information that can be extracted from the dataset underlying the Topic but cannot be represented on a graph.

Therefore Crystal doesn't just answer to your question providing punctual graphic information, but also helps you understand them better providing a complete and accurate description of what data are actually telling below the surface.

In particular, Automated Insights are based on a Data Analysis technique called Descriptive Analytics.

Pills of Data Analysis

Descriptive Analytics are called like this because they describe the dataset, especially focusing on what has happened in the past and on what is happening at present.

Discover The Hidden Insights

Accessing the Automated Insights of your Topics is straightforward:

  1. first of all, make your question to Crystal

  2. notice the Right Sidebar of the Advisor Workspace

  3. there, you will find the list of all the Insights available for your Topic

At this point, you can scroll through the Insights provided and let them guide your analysis: the information provided will help you understand better which are the most relevant aspects of the Topic and how to continue your analysis.

Please Note

  • The number of available Insights for each Topic can change from a minimum of one Insight to twelve Insights maximum, based on the Topic's underlying data

  • Insights are dynamic: since they are always related to the Topic you just asked, they will always change based on your questions

How Do They Work?

As Descriptive Analytics, the Automated Insight can shed light of a number of information related to what happened to your data in the past and what is happening right now.

The number of Insights available for a Topic depends on the type of Objective and Visualization, because each Objective / Visualization has specific information that can be extracted from it.

However, the actual number of Insights shown for a requested Topic depends on the specific requirements needed to calculate each Insights.

Therefore, whenever a specific Topic is asked, Crystal detects the type of Objective / Visualization and, from the list of available Insights linked to it, tries to calculate as more applicable Insights as possible based on the specific data underlying the Topic.

The final result of displayed Insights can be a number between 1 and 12 Insights.

Insights Catalog

Following, you can find the complete list of the Automated Insights that Crystal is able to extract.

Please Note

This list reports the general purpose and logics of the analysis done by each Insight.

However, in practice, the specific calculations done vary slightly to adapt to the Objective / Visualization they are applied to.

N.InsightLogic

1

Average and Variability

Calculates the Mean and the Standard Deviation of an Entity.

2

Major Mean

Calculates the Entity with the highest average of all entities.

3

Minor Mean

Calculates the Entity with the lowest average of all entities.

4

Highest Sum

Calculates the total sum of the relevant values associated with a specific Entity. It then groups these values according to a category and identifies which category has the highest total sum, thus highlighting the Entity with the highest value within the dataset.

5

Lowest Sum

Calculates the total sum of the relevant values associated with a specific Entity. It then groups these values according to a category and identifies which category has the lowest total sum, thus highlighting the entity with the highest value within the dataset.

6

Extreme Values

Calculates the Minimum Value(s) and the Maximum Value(s) of an Entity.

7

Unique Values

Calculates the number of distinct values of an Entity.

8

Focus on Last Timeframe

Calculates the Entity Value at the beginning and at the end of the last consolidated timeframe, based on the Time Aggregation configured for the Topic (Months, Quarters, Years, ...).

Calculates the Average of the Entity across the Time Range.

9

Global Change

Identifies the Entity Value at the beginning and at the end of a Time Series, based on the Topic’s Time Range.

Calculates the difference between the starting point and the end point of the Time Series.

Calculates the percentage difference between the starting point and the end point.

10

Highest Difference

Calculates the highest difference, both in absolute value and in percentage, among consecutive values.

11

Lowest Difference

Calculates the lowest difference, both in absolute value and in percentage, among consecutive values.

12

Interesting Event

Searches for statistical anomalies.

If multiple anomalies are identified, selects the most “anomalous” anomaly.

13

Timeframe Comparison

Compares the values of on Entity in two different time periods, then highlights the variation.

14

Top Contributor

Ranks all Entity Values from highest to lowest.

Identifies the one with the highest value.

Calculates the percentage of this highest value against the Total Sum of the Entity across all Values.

15

Bottom Contributor

Ranks all values of the Entity from highest to lowest. Identifies the one with the lowest value. Calculates the percentage of this lowest value with respect to the Entity's total sum of all values.

16

Best Performer

Groups by one category, then identifies an entity that is first for an entity in the other category and at least third for another entity.

17

Worst Performer

Groups by one category, then identifies an entity that is last for an entity in the other category and at least third last for another entity.

18

Percentage Difference

Calculates the difference between two percentages.

19

Cluster Comparison

Finds the total average, then the average for the top N, and the average for the bottom N.

20

Distribution of Data

Calculates how many positive and negative values are present.

21

Mean of Positive and Negative Values

Calculates the mean of positive and negative values.

22

Total of Positive and Negative Values

Sums the positive and negative values.

23

Correlation

Calculates the correlation between two variables.

24

Highest Similarity

Calculates how many times the time series overlap with each other.

25

Value of Progress

Highlights the value of the progress towards a target.

26

Percentage of Progress

Highlights the percentage of target reached.

27

Most Common Value

Highlights the most common value found.

Keep analyzing your data!


Last updated