Last updated
Last updated
Before you start talking with your data, it is crucial for you to familiarize yourself with the concept of what we call "Topic".
Topics are the foundational units of information that enable meaningful interactions with Crystal.
In fact, asking Crystal a question means asking about a Topic, and at the same time, receiving an answer means receiving a Topic.
Remember
In order to be able to ask about and read Topics, you first need to understand what Topics are made of.
So, let's get started!
From an analytical perspective, Topics are aggregations of data that form a specific Key Performance Indicator (KPI) or Insight. As such, they represent a specific data analysis.
These are the data types that constitute a Topic:
| N. | Data Type | Mandatory / Optional |
|---|---|---|
1 | Objective / Visualization | Mandatory |
2 | Topic Chart | Mandatory |
3 | Main Entity / Main Entities | Mandatory |
4 | Time Range Filter | Mandatory |
5 | Dynamic Filters | Optional |
6 | Key Values | Optional |
Let's dive deeper into each of them!
The Objective is the purpose of the data analysis expressed by the Topic.
The Objective answers the question: "What specific insights or outcomes am I seeking with this particular Topic?"
Each Objective is further illustrated by a specific Visualization, i.e. a specific visual support to express the Objective. Each Objective can be expressed by one or more Visualizations.
Discover the Objectives available on Crystal, their purpose, and their Visualizations:
Remember
You can introduce a Topic with:
an Objective name (e.g. Composition)
a Visualization name (e.g. Breakdown)
a specific formulation (e.g. The Total of...)
All these formulations express the purpose of the Topic.
You can dive deeper into Objectives and Visualizations here.
The Topic Chart is the actual graphical representation of data based on the specific Objective and Visualization expressed by the Topic.
Each Visualization has its own Charts, for example:
Comparisons can be represented by Bar Charts or Scatterplots of different kinds
Progress can be represented by Gauges of different kinds
Breakdown can be represented by a Donut Chart
Please Note
Topic Charts are always interactive, allowing you to explore your data in greater depth!
For example, you can engage with the legend to filter out some data or you can highlight parts of the Chart to access detailed information via a tooltip.
Remember
Entities are the core elements of the Topics.
An Entity denotes a singular unit of data, usually a dimension or a metric.
Examples of Entities are: "Sales", "Cost", "Target", "Country by Product Name", "Average Sales", etc.
Usually, an Entity can be custom-created using formulas, facilitating the aggregation of data through specific computations (e.g., count, sum, average).
When an Entity assumes the role of primary metric within the Topic, it is called a "Main Entity".
Topics can have a single main Entity or more main Entities, based on the type of Objective and Visualization. For example:
a Composition Topic is made up by a Single Main Entity
a Multi-Line Graph can be made up by Multi Entities
A Time Range Filter is a specific time restriction applied to a Topic to limit the scope of the analysis.
Please Note
Each Topic always comes with a Time Filter applied by default, because it is not possible to have an analysis that is not time restricted.
Dynamic Filters are Entities used to narrow down the Topic's analysis scope, by filtering the main Entity based on a specific Value.
When an Entity takes on the role of a filter, like in the case of Dynamic Filters, it is called a "Filter Entity". Dynamic Filters can hence also be called "Filter Entities".
Remember
To sum up, there are two different kinds of Entities possible on Crystal:
Main Entities (the subject of a Topic)
Filter Entities (the filters applied to a Topic)
Dynamic Filters can be:
Numeric, either positive or negative
For example: currencies, item codes, sizes, etc.
Non-numeric
For example: countries, towns, areas, item names, customer names, etc.
Each Filter Entity is defined by a Value (also called Entity Value): the elemental data units that define the specific metric represented by the Filter. While an Entity represents a metric, the Entity Value is the concrete instance or measurement within that metric.
For example:
If the Filter Entity is "Revenue”, the corresponding values could include specific amounts like "$10,000," "$15,000," etc.
If the Filter Entity is "Country", the corresponding values could include "Argentina", "Italy", "United States", etc.
Please Note
The presence of Dynamic Filters is optional.
Key Values are numeric values that summarize data relevant to the Topic.
Please Note
The presence of Key Values is optional.
Remember that each Topic always comes inside of a Topic Card, which acts like the Topic’s control panel.
Here are the components of a Topic Card:
Topic Title The summary of the Objective and Main Entities
Topic The actual Topic with all its components (Filters, Chart, Key Values)
Topic Actions All the actions that can be performed on the Topic or the Topic Card
Here is a list of all the Topic actions you can perform from the Topic Card, based on its access point.
For more information about each capability, please refer to the dedicated section of the documentation.
At the bottom of the Topic Card, you'll find a section of useful functions available with just one click.
These functions include:
At the top right of the Topic Card, you'll find a dedicated menu - with three dots - to access useful functions with just one click.
These functions include:
Since Release 2.58, at the top right of the Topic Card, next to the menu, you'll find a new button to change Chart in just one click.
You can choose to view a Topic Card in expanded mode. Just click on the "Expand" button next to the menu on the top right corner to see a bigger version of the Topic Card and explore the Topic Chart more easily.
The expanded Topic Card still carries all the actions and Insights available in the Card's normal version.
All the Topics available on Crystal look the same, are made of the same elements, and support the same actions from the Topic Cards.
However, the recently introduced Entity-Based Conversation capability has defined two different types of Topics to consider:
Configured Topics
Auto-generated Topics
Configured Topics are the results of the Topic Configuration process, also called Topic creation, which has been the primary way to generate Topics until now.
Topic configuration / creation can only be implemented by Admin Users in the Crystal Console and basically consists in manually creating the Topic from scratch by choosing all the elements described in the sections above (Objective, Entities, Filters, etc.).
These kinds of Topics will be available for Member Users in Advisor only after the conclusion of the creation process (publication step) and only if the Admin has granted the viewing permissions for those Topics.
Here are a few simple conversational rules to ask about this kind of topic.
Please Note
When interacting with the configured Topics in Crystal Advisor, some of the elements chosen by the Admin will remain set and not editable, whereas others can be changed.
Non-editable elements:
Objective and Visualization
Topic Chart
Main Entities
Key Values
Editable elements:
Time Range Filters
Dynamic Filters
In particular, Filters in the Advisor can be changed in several different ways: either in aConversation, through the Filter Topics capability, or in the Topics Section.
As the name suggests, Topics that are auto-generated are automatically produced by Crystal, without the need for the Admin User to create them from scratch.
This is a new method for Topic generation, which allows you to ask questions beyond the scope of Configured Topics, i.e. you will be able to ask about any Topic, without necessarily relying on Topic Creation.
You can generally ask about auto-generated Topics the same way you ask about configured Topics.
Take a look at the "Entity-Based Conversation" article to learn more about how Crystal generates Topics, and how to ask about them.
Now you are ready to explore your data!
