As a follow-up to my last post, when dashboards first came up, I thought it would be useful to explore futher not only the dashboard itself, but also about the graphs that fit into this type of information medium. Before we move on to the graphs themselves, let's outline a broader context for using dashboards.
In many respects, the dashboard is a new version of the Executive Information System (EIS), whose task is to facilitate and support the collection of information (from both external and internal sources) that is helpful in decision-making relevant to the goals and expectations of the organization.[1] In order for a dashboard to perform this task well, any graphs used on it should not only be informative, but also clear. As Stephen Few writes, the basic challenge in creating a dashboard is to place a lot of information in a small space so that everything is understood at first glance.[2] But is that the only challenge?
What are the characteristics of a good dashboard? Here are the 7 basic principles:
- It shows the information necessary to achieve a specific, set objective - not more and not less (these may be KPIs, but this is not a prerequisite: the information is to be useful).
- It fits the computer / device screen: the whole dashboard is visible at a glance without scrolling.
- It is dominated by graphic objects; text is a possible addition with the caveat that a well-designed graphic design conveys more information than text.
- The information is presented in summary form, in the form of summaries or exceptions. A dashboard is used to monitor a situation when a busy executive is under time pressure, so the presentation of information must be well thought out so that the graphics make decision-making easier, instead of focusing on irrelevant details.
- Contrary to a popular belief, it does not have to be presented using a web browser (although currently it is one of the best media), or refreshed in real time (it is important, however, that the data are refreshed according to the specificity of the topic, and that refreshing is possible and easy to perform).
- The dashboard is adjusted to the requirements of the recipient.
- Objects conveying information are clear, concise and intuitive: they clearly define the content of the message without taking up a lot of space. Hence, the frequent use of gauges (like fuel gauges), thermometers, bricks, and traffic lights.[3]
Given the above requirements, it is easy to see why the bullet graph discussed in my last post is a suitable graphic to be used on the dashboard. Another visualization that lends itself well to dashboards is the thermometer graph. Let's take a look at the chart below.
We return again to the subject of smog. An important determinant of air quality is the concentration of pollutants, as already discussed in an earlier article. Equally important is the composition of pollutants and how they change over the year. PM10 dust is made up of metals and benzenes, which are the result of e.g. combustion of unsuitable materials in furnaces. Let's see how the composition of dust changes over the year. The benzenes are marked with a dark grey bar on the visualization.
The visualization uses average monthly data from the Małopolska region for 2016.[4] Data presented in the form of a thermometer makes it easier to communicate information: everyone knows how to interpret it because everyone knows how to read a thermometer. With this clear form with no unnecessary embellishments, and with good use of space, we can read the data very easily:
- the average monthly concentration of benzenes in PM10 dust in Małopolska in 2016 was highest in February: benzenes accounted for 74% of the dust content;
- the lowest average monthly concentration of benzenes in PM10 dust in Małopolska in 2016 was 11% in July;
- it is clearly visible that the average monthly concentration of benzenes in PM10 dust in Małopolska in 2016 was growing in the so-called heating season, i.e. in the 1st and 4th quarter of the year.
In conclusion, the thermometer graph, combined with the bullet graph from the last post, provides us with two graphs that readily lend themselves to the creation of effective dashboards.
The thermometer graph is one of 50+ custom built procedures included with PS IMAGO PRO.
[1] The definition after Wikipedia, access: 31.10.2017. [2] See: Stephen Few, Dashboard Confusion, Intelligent Enterprise, 2004; or: Information, dashboard, design. The effective visual communication of data; O'Reily Media, 2006. [3] See: Stephen Few, Information, dashboard, design. The effective visual communication of data; O'Reily Media, 2006. [4] Data after: http://www.gios.gov.pl/pl/