{"id":19920,"date":"2022-11-21T15:13:12","date_gmt":"2022-11-21T14:13:12","guid":{"rendered":"https:\/\/www.graphomate.com\/2022\/11\/how-to-show-trends-von-sparklines-wuermlis-und-spaghetti\/"},"modified":"2022-11-28T13:59:18","modified_gmt":"2022-11-28T12:59:18","slug":"how-to-show-trends-von-sparklines-wuermlis-und-spaghetti","status":"publish","type":"post","link":"https:\/\/www.graphomate.com\/en\/2022\/11\/how-to-show-trends-von-sparklines-wuermlis-und-spaghetti\/","title":{"rendered":"How to show trends – Sparklines, Worms and Spaghetti"},"content":{"rendered":"\n

I\u2019ve been a big fan of sparklines for over 15 years now, so I thought it was high time I provided a more in-depth review of these \u201cword graphics\u201d.<\/p>\n\n\n\n

The \u201cinvention\u201d of sparklines is attributed to Edward Tufte, as briefly mentioned in a previous post<\/a><\/strong> of mine. In his 2006 book Beautiful Evidence<\/a><\/strong>, he describes sparklines as \u201cintense, simple, word-sized graphics that depict trends in a meaningful and information-dense way.\u201d Here is an example of such a sparkline:<\/p>\n\n\n

\n
\"Liniengrafik<\/a>
Sparklines by Edward Tufte from 2006<\/figcaption><\/figure>\n<\/div>\n\n\n

How are sparklines used on dashboards?<\/h2>\n\n\n\n

Back in 2008, I built a dynamic dashboard [sic<\/em>.\ud83d\ude09] in Excel that made extensive use of sparklines. Here is a section from that dashboard:<\/p>\n\n\n

\n
\"Gewinn<\/a>
Dynamic dashboard with sparklines from 2008<\/figcaption><\/figure>\n<\/div>\n\n\n

On the far left, the changes in five key figures over 12 months are shown with planned values in red.
The target achievement for the entire year is shown with bullet graphs<\/a><\/strong>, a visualisation alternative to Stephen Few’s speedometers.
In finance and controlling, this combination of trend display and the display of the degree of plan fulfilment for the year to date is common convention.<\/p>\n\n\n\n

But are sparklines easy to interpret?<\/h2>\n\n\n\n

Sparklines show the trends of several time series in a very small space in accordance with the spirit of Tufte: \u201cAbove all else, always show the data.\u201d We can, at a glance, see patterns and outliers in the lines without delving into too much of the detail.
This is also in the spirit of IBCS (International Business Communication Standards) with its demand for high information density in the visualisation of information: Condense<\/a>.<\/p>\n\n\n\n

But for Rolf Hichert – the founding father of IBCS – sparklines are a no-go: “I am very sceptical about the use of sparklines in dashboards”<\/a>. We\u2019ve even had our fair share of professional disagreements about the pros and cons of these trend displays since he and I first met in 2009.<\/p>\n\n\n\n

But why exactly does Hichert never tire of dismissing sparklines somewhat disrespectfully as “worms” in his presentations? The reason is because they are not uniformly scaled<\/strong> and thus not visually comparable to each other.<\/p>\n\n\n\n

Here is an example from Hichert’s presentation Small is Important, But Not Always Right<\/a> to illustrate his repugnance towards sparklines. 240 values of a table are first visualised as individually scaled sparklines and for comparison by means of a uniformly scaled line chart.<\/p>\n\n\n

\n
\"\"<\/a>
example by Rolf Hichert<\/figcaption><\/figure>\n<\/div>\n\n\n

Let us look at the first two sparklines A + B highlighted in yellow, whose course is identical. In the “spaghetti lines” to the right of them, it becomes clear that data series B is to be evaluated quite differently from series A. The data points are smaller by a factor of about 30.<\/p>\n\n\n\n

From the above, it follows that the dips in the middle segment of sparklines A + B are also to be evaluated very differently i.e., B is probably negligible. Nevertheless, a pattern is clearly recognisable <\/strong>which is almost completely impossible with the spaghetti lines! To analyse the trends in more detail, users can be given the possibility to use further tools such as filters, jumps, tool tips etc.<\/p>\n\n\n\n

Building sparklines with better UX and interpretability<\/h2>\n\n\n\n

During the development of our graphomate sparklines for SAP Lumira Designer<\/strong> we came up with some ideas to support users in their correct interpretation as illustrated in this two graphics.<\/p>\n\n\n\n

\n
\n
\"\"<\/a>
graphomate sparklines for SAP Lumira Designer<\/figcaption><\/figure>\n<\/div>\n\n\n\n
\n
\"Sparklines\"<\/a>
graphomate sparklines -pop-up window<\/figcaption><\/figure>\n<\/div><\/div>\n<\/div>\n\n\n\n

In particular, we decided to harness the interactive capabilities of dashboards, which gave us these functionalities:<\/p>\n\n\n\n