You are entering the Business Intelligence world and you do not know where to start? DO not panic, our guide will help you assimilate the BI base vocabulary.
Business Intelligence, dashboards, reporting… it has already been a long time since we started discussing these terms. Each Monday, we were struggling to find a simple way to explain some Business Intelligence vocabulary. It is now done via this article.
dimensions, measures, hierarchies… What would you say about a short article explaining the base vocabulary?
You are more numerous wanting to use dynamic dashboards with the goal to automize your tasks and to gain time.
In order to help you and to help the business user (who is not supposed to be an IT professional), we want to go back to the basic vocabulary that surrounds the Business Intelligence world and therefore help you to start on strong bases.
Definition n°1 (and the most important): What is Business Intelligence?
Business Intelligence is the computing that allows a company’s collaborators to visualize data to make the correct decisions at the right moment.
Definition n°2: What is a measure?
BI’s base, a measure is a numeric indicator that represents a size (the revenue or a product’s price, for example). The mnemonic way to remember it is to think that a measure can be measured.
Definition n°3: What is a dimension?
A dimension is a set of elements that can be organized or not according to one or more hierarchies. A dimension can be discrete, which implies that its elements do not have any particular order (for example, French departments). it can also be continuous, in that case, the order of its members has importance: the dimension of time.
Definition n°4: What is a hierarchy?
A hierarchy is a logical organization of elements of a dimension in a hierarchical manner. For example, a geographic hierarchy which the highest level would be the continent would include the level “country”, and the level “region” and so on…
Definition n°5: What is aggregation?
Aggregation is the result of a value resulting from a calculation (sum, average, min, etc) to a measure explored by one or more dimensions. For example, the sum of revenue calculation of a store during a trimester is the result (aggregation) of the sum of a daily revenue data set during a trimester.
Definition n°6: What is a data cube?
A data cube is a method to store data that contain the definition of measures, dimensions, and hierarchies. This structure is used in BI software and allow to rapidly and simply obtain aggregated data.
Definition n°7: What is a dynamic dashboard?
A dynamic dashboard is a set of graphic representations on which the business users can not only visualize and interact with the relative information at their availability but also communicate with their collaborators in live time.
Definition n°8: What is a key performance indicator (KPI)?
It is difficult to discuss dynamic dashboards without mentioning indicators. Key Performance Indicators are the indicators on which a company’s collaborators will base themselves in order to know which objectives have been attained. We often associate an objective to reach and we can color code to qualify the completeness of the objective (green, yellow, red).
Definition n°9: What is data visualization?
Data visualization is the art of using the appropriate graphs or animations to present data or indicators visually. The objective is to use the correct graph at the correct time, to transmit the appropriate information to the user. You would use column graphs if you group a measure according to a continuous dimension, whereas in the case of a discrete dimension it is more adapted to use a horizontal bar graph.
Definition n°10: What are a data mart and a data warehouse?
A data mart is an organized data warehouse that allows responding to the problematics encountered by a company’s business unit. A data warehouse is a set of data marts. for example, if a company has various departments (marketing, finance, HR), we would have three data marts associated with those departments. The grouping of these data marts is called a data warehouse. The following example has for an objective to put in context the entirety of the definitions we have presented.
|Date||Continuous dimension||Date when the data has been recorded|
|Region||Discrete dimension||Geographic region where the data has been recorded|
|Business Unit||Discrete dimension||Company’s business unit (HR, Marketing, etc)|
|Material||Discrete dimension||Phone model|
|Length of communication||Measure||Time in minutes|
We have four dimensions of different natures (discrete and continuous). The dashboard software DigDash Enterprise allows going further by automatically adding two other types of dimensions
- A time dimension (date)
- Geographic dimensions (region)
It brings us directly to the hierarchy notion. When DigDash Enterprise detects a geographic or a time dimension, it automatically creates the following hierarchies:
It is also possible to manually create your own hierarchies. For example, in the case of elements, it could consist to of classifying the different phone models in two categories:
- Mobile phones
To better illustrate the previous points as well as the aggregation and the data cube notions, we will go over the data visualization portion. We will build a map on which we wish to visualize the average communication quality according to geographic zones.
If we position ourselves at the highest geographic hierarchy (continent), we obtain the following map:
In this case, the data has been aggregated at a continental level. It means that we have placed the average of the American continent on the map, as well as the average of the European continent.
We also have the possibility to navigate through the various geographic hierarchies by clicking on the map. By clicking on Europe and then on France, we obtain a new map:
This time, the data has been aggregated at a regional level and not continental. The data visualization at a regional level is almost instantaneous. This is possible thanks to the data storing technology within the DigDash cubes. It also allows rapid navigation, even on a large amount of data, o the contrary of databases.
30 January 2019