Definition and the rise of dashboards
A dashboard refers to a graphical user interface for the visualisation of data. The word dashboard has mostly been associated with cars and in that context it denotes a panel with gauges and controls. As is generally known, this type of dashboard displays not only the speed but also warnings from the engine compartment and early warning indicators from driver assistance systems. The use of such powerful dashboards has now become standard practice in many companies and departments. However, studies have shown that businesses with up to 1,000 employees still have an enormous need to catch up in this respect.
The benefit – Added value can be realised
The added value of dashboards is that they make it possible to visualise numbers in very many different ways. This visualisation means that the report recipients are provided with a completely new approach to the data and its interrelationships. The causes and effects of various metrics and developments can thus be intuitively understood and experienced.
Besides the well-known visualisation options such as line charts and bar charts, there are, in particular, so-called heat maps or map sections where the data can be shown on the basis of the regional distribution (e.g., of customers or products).
Here the graphics can be designed to be interactive. By clicking on the graphics it is possible to pull up details; for example, by clicking on the sales revenues it would thus be possible to display the allocation between the product categories. By continuing to click you could get information on margins and customer groups.
Advantages of using dashboards
The potential applications of dashboards have increased considerably in recent years and we have an overview of these below (cf. also the sample graphics).
- All the graphics can be customised and called up on tablet computers or mobile phones;
- a data connection to the ERP system is possible;
- the graphics can be designed to be interactive and you can get a feel for the details simply by clicking;
- the integration of artificial intelligence makes it pos-sible to incorporate algorithms to forecast metrics related to sales and costs;
- scenario programming allows answers to be provided to an array of questions, e.g.: What is the outlook for the development of the situation of the business for the coming months if the borrowing rates go up by x% and/or the purchase prices by y% or if sales deteriorate by z%?
Project sequence for the introduction of dashboards
Definition of project goals
At the start of the project, you need an answer to the question of where the dashboards should be deployed. We would recommend that you start with one department so that you are able to transfer the experiences and benefits more quickly to other departments. Dashboards are particularly suited for use in the following departments:
- finance and accounting,
- controlling (at the central and departmental levels),
- personnel department,
- production planning and control,
Besides the selection of the department where the dashboard should be deployed, the information needs of the report recipients are also particularly important. These are normally based on the previous reporting and include the following data:
- historic developments,
- measurement and comparison of developments over time,
- comparisons between geographically different courses of development,
- price differences.
Besides the question of ‘what’, asking ‘how’ is also particularly crucial. And this is where the data scientists are deployed. The visualisation should indeed be optically appealing, however, the actual benefit only arises if the graphic contributes to the informative value and the interpretation of the data. A bar chart with postcodes for the analysis of the structure of the customer base could be good; however, it would be better to underpin the analysis by displaying a map and using this to visually represent the regional distribution of customers.
Once the goal has been defined, the next step involves clarifying whether or not the required data are available in the ERP system or in previous/feeder systems and whether or not these systems can be accessed via an interface (API). It is likewise important to clarify how frequently the report will be called up and how up-to-date the report needs to be. A distinction thus has to be made in terms of whether it should be a data model where the visualisation would be on the basis of daily updates, if possible, (up to ‘near real time’), or whether weekly or monthly reports would be sufficient.
Once the specific informative value for the report recipients has been defined, the data model and its granularity have to be reviewed. It is frequently ascertained that either even more details would be possible or new data sources need to be opened up. However, caution is required for this step because dashboards are not databases. The more complex the data model and the calculations, the slower the performance becomes. That is why larger datasets should be mapped directly via a database interface, or large-scale calculations via software such as Python.
Existing reports and reviews in all departments are predestined to be switched to modern dashboards. Using visualisation and an interactive presentation is a considerably better way to reach target readers and, in turn, they will be able to derive decisions with a focused perspective. Good preparation and a positive introduction in all the departments will allow markedly faster decision-making while, nevertheless, maintaining accuracy. Data scientists can provide support when designing complex applications.
Please note: More information at: www.PowerBWA.de