Analytical desktop tools

Analytical desktop tools can be described as a new generation of reporting tools. They weave a transparent layer over the database structure and thus offer various ad hoc analysis options from the desktop.

Ad hoc

Although the use of an analytical desktop tool is not recommended at OLTP level due to the risk of performance problems, it comes into its own in a structured information environment. We can even say that the information environment loses importance in its absence. In the case of an ODS, one could still somewhat question its usefulness, given the operational nature of such an environment. However, since the amount of data itself is limited, the main reason for not using it disappears. Nevertheless, the very best use lies in the approach of a dimensional data model). The symantic layer of such applications uses so-called objects , where each object theoretically corresponds to a column in a database table. When building the layer, one can in a sense adopt the star schema of a data warehouse and implement the same logic 1 on 1. Within the object structure, facts and dimensions can be presented separately and the business user can, as it were, look inside the database, albeit in a user-friendly way. To clarify this, let’s take the example of a setup within the Business Objects tool. The symantic layer is called a universe within this software . The figure illustrates how the business user can be presented with the database structure of a financial data warehouse.

Business Objects example – universe” onclick=”openImage(this);”> Business Objects example – universe

The structure is divided into objects, which are placed in so-called classes. Filing offers the opportunity to make the presentation slightly different from the database structure and thus increase transparency. All classes that refer to codes in their names in the example above can be considered dimensions, while the other classes will include the fact data. When discussing a dimensional data model, it was mentioned that a dimension within a report can serve as a selection dimension and as a criterion dimension. This is also clearly visible in the figure, where the selection dimensions will be selected for the result objects, while the criterion dimensions will be used for the conditions. The figure shows the result of a query constructed based on this universe. The result objects are displayed on the left side of the screen, the actual information is shown on the right side.

Business Objects example – result” onclick=”openImage(this);”> Business Objects example – result

It is of course not the intention to explain how Business Objects works. The relevance of the above examples lies in the fact that this tool is currently the market leader in the field of analytical reporting. As a result, the illustrations provide a clear picture of the general operation of such applications and the direction in which they are evolving.

Data Cube

One of the most important features of a good analysis tool is the presence of an internal data cube , which will store the extracted data internally for further use within the application. Depending on the number of selected dimensions, it can range from a one-dimensional to a multi-dimensional cube. Since almost every report uses at least a normal dimension and a time dimension, we can assume that we are always dealing with the second scenario. The advantage of the data cube is the separation of extraction and use. For example, a report can be scheduled to fill the internal cube with a large amount of data at night, so that this data can be used immediately in the morning to build a representative report. As is often the case, this advantage is also the biggest disadvantage. Because the data is stored internally within a report that is stored locally on the file system, a transformation takes place, as it were, of central data into local data. If the data within the local cube is not refreshed on a regular basis, there is a chance that at some point an inconsistency will arise between the data in the data warehouse and that within the report. One notices that the structure of a data cube is somewhat similar to that of a MOLAP cube, with the difference that the MOLAP cube is located centrally on the database and the data cube is a local data item.

Web analysis tools

The Internet has made the world smaller and the need has grown for tools that can display the necessary information at any time and from any location. That is why most desktop tools nowadays provide a web-based equivalent. People are sometimes somewhat skeptical about the tendency that certain vendors have to also provide every functionality of their desktop tool in a web application. It causes unnecessary complexity and a loss of operational power within a platform that is not intended for this.

Business Objects example – web

A desktop tool is more suitable for building complex reports with a high level of presentation, while the web emphasizes quick access to information, whether or not via ad hoc techniques. Combining both within the same web application seems to have little added value and can only be detrimental to the ultimate intention of using such an application. The figure provides an illustration of the Business Objects dashboard application.