When considering the development of a efficient dashboard we must take into account a number of basic characteristics. These points are very important for the development and consumption of information. We can divide them into:
- Coherence of data.
- Organisation and simplicity.
Consistency of data on a dashboard
Coherent information can be obtained by answering several basic questions.
What do I want to show?
It is important to know what kind of information we are going to show in our reports. Generally, this information is provided by the user who will be consuming the dashboard. Sometimes, users may not know what is possible or what is not possible to represent because they are not familiar with the capabilities of the visualisation tools. In these cases, help and advice from developers is vital.
Who is the information I am going to represent intended for?
Another aspect to take into account is to know who will be the final consumer(s) who will exploit this information. It may be that the client/user with whom we have interacted the most for the construction of the report will not be the only one who will access the data. For this reason, it is essential to be transparent and decide how best to represent that information.
How should I represent this information?
In line with the previous point, once we know to whom the information we are going to show is addressed, we have to see how to do it in the best way.
Typically, end users who are going to exploit this data do not want a lot of hassle when navigating the dashboard. Therefore, it is key to build a comfortable and 'user friendly' navigability to reduce the complexity when exploiting the information.
Organisation of the dashboard
When we talk about organisation, we understand it as the location of visual objects in the report. Obviously, it must also have a coherent structure. For example, we can divide the dashboard into three sections: top, middle and bottom.
At the top we will place high-level insights such as KPI's, titles, logos, etc. The idea is to use only a 30% of the space for this purpose.
In the centre, there can be visuals that represent the trend of the data or metrics based on the activity we want to represent: tables; large visuals such as bar charts, representative maps, etc. occupying 50% of the space. The lower section is used to represent granular metrics and very specific KPIs or tables. This section usually occupies 20 % of the dashboard.
The representation of the data can be reflected in six categories:
Comparison: Compare data between different categories.
- Most commonly used visual objects: grouped and ungrouped bar charts, line charts, stacketed charts, bubble charts, etc.
Data over time: Represent the trend and changes in data over time.
- Most commonly used visual objects: bar chart, line chart, area chart and waterfall chart.
Correlation: Visualise the relationship between two or more variables.
- Most commonly used visual objects: bubble chart, column chart, line chart and dot chart.
Distribution: Visualise how the data occurs and is distributed in our dataset.
- Most commonly used visual objects: Histograms or clustered bars.
Part/whole relationship: Show how certain elements form part of a whole.
- Most commonly used visual objects: Clustered bars or Treemap.
Ranking: Display the position of the elements in their order of importance.
- Most commonly used visual objects: Ordered columns or bars, funnel charts, etc.
Of all the charts, one of the most widely used are pie charts and derivatives which, although they are very popular, there is an implicit rule (within good practice) in terms of visualisation: whenever possible, we should avoid using pie charts. They are not graphs that provide more or better information than the other alternatives and can even be confusing if there are categories with similar frequencies.
Clean and simple dashboards
Cleaning is understood as the action that consists of not saturating the dashboard with visualisations, objects, tables, etc.
Reports are intended to make information comprehensible and understandable, as well as to facilitate decision-making. Too many visual objects can lead to the opposite, creating confusion and worsening the performance of the report, which can frustrate users.
By simplicity we mean a report that does not use too many advanced features. The most common type of report usually represents static information with little user interaction (beyond being able to filter and select key aspects to exploit the data).
However, there are times when it is necessary to extend the complexity of the report by adding features such as Drill Down or Drill Throw, which allows for exploration of the granularity of the data down to the detail, or buttons that trigger a particular action.