Let me guide your series: How to create a top-notch dashboard with self-service BI

Geschreven door Iver van de Zand op


Let me introduce you to the first episode of a series of ‘Let Me Guide You’ articles in which I’ll share my experience and best practices on data visualizations. This one is about creating top-notch stories and dashboards using self-service BI. Select the appropriate BI components and off you go!

Infographics and storyboards act like the icing on the cake. When you have finished your data explorations and data insights, you’ll decide what type of visualization to use to reflect your insights in the best possible way. We discussed this in an earlier blog. You also have to choose what visualizations to bring onto the storyboard and how to structure and layout them. The key questions here are:

  • What exactly is the insight I want to communicate? The ‘exactly’ refers to your target audience (think about the level of abstraction), your perspectives of the insight (time-series, GEO) and your story flow.
  • What is the experience I want my user to have? Is your story dashboard-like, i.e. indicating KPIs and trends on a regular basis and do you want to re-use it? Do you want to offer the user more detail, i.e. does your story require multiple pages with summarized and detailed insights?  Or do you want to deliver an infographic, which provides context-sensitive visualizations? Does your story need to be interactive?

So let’s go and find out some of the better practices to create impactful boards. In this series we look at KPI-driven dashboards. I use SAP Lumira since it is one of the most flexible tools for agile data visualization and one of the easiest, interactive self-service BI tools available in the market.

Hichert standards

pic2For KPI-driven dashboards I’d recommend working to the International Business Communication Standards (IBCS) where possible, as they greatly help communicate insights in an efficient, rational way. IBCS uses the standards once developed by Dr. Rolf Hichert (Look it up: the ‘Hichert Success Rules’). Applying – pieces of – these standards for your generic BI design certainly pays off; it will help users to adopt and respect your BI insights, thus minimizing the chance they’ll misinterpret these.

IBCS offers a whole variety of communication standards; advising what chart/graphs to use when, how to deal with captions, axes, color coding and much more. SAP Design Studio and Lumira graphomate charts, for example, are `Hichert-certified’.


Landing pages and font types

Using landing pages as a front page for your story or dashboard, is a great way to welcome your users and guide or navigate them through your story. On the landing page you can create categories of information that users can link to. By clicking one of the categories, they’ll go to the correct section or page of your board. Use landing pages if:

  • you want to split the summarized KPIs from the detailed ones. A user might only want to see the detailed information of a KPI once the KPI is behaving inconsistently.
  • you want to communicate insights on different information domains, i.e. if you create a dashboard with one overview KPI page describing ‘HR Attrition and HR recruitment metrics’, ‘Vendor statistics’ and ‘Marketing lead generation’, you might want to included details for these domains on separate pages. The landing page will provide input to the applicable domains.

Landing pages serve different purposes:

  • They welcome the user and allow of text/video communication explaining what the storyboard is all about, how it should be used and what to do in case of remarks.
  • They act as a navigation guidance tool through your story. On a landing page you can create links towards other sections/pages in your story.
  • Landing pages guide users to additional information outside your story. On a landing page you can create URL links to any http address.
  • Landing pages can serve to set data filters that need to apply to all pages and visualizations in your board. For example, if a user wants to study all the information on your storyboard only for the years 2013 – 2014 – 2015, he can apply these filters on your landing page instead of doing that page per page.


Color-coded performance indicators

When I create storyboards that require color-coded bars to indicate a KPI trend or KPI performance, I use conditional formatting as the base. This is how it works step-by-step:

  1. Create a crosstab with your KPI metrics and applicable dimensions and attributes.
  2. Create conditional formatting on your metric by applying rules and color-coding.
  3. Change the crosstab into a bar or pie chart.
  4. Use the applicable dimension as a filter to apply the KPI color-coded bar in detail.
  5. Choose a picture that applies to the subject of your insight and apply it to your storyboard. In the video I use a glass to indicate how much it is `filled’ with the KPI performance. An empty glass is a low performing KPI …
  6. Drag and drop your color-coded visualization onto your storyboard.
  7. Now when changing filters or range, the color-coded bar will automatically adjust.

pic3Enrich the above with small (!) text comments for KPI labeling. Try to avoid bold font types – they deflect people’s attention from the graphs – and use regular fonts instead. A font type that works very well is the modern-looking Trebuchet. The color of the font should not be black but dark grey; again this gives a more modern look.

Never use more than three to four colors in total on your KPI overview dashboard. Two colors are used for both performing and underperforming metrics, the third is used as background and the fourth can eventually be used as plot background for your chart. Commenting, illustrating and text boxes always use one of the four applicable colors.



Enriching visualizations with baselines

Especially in situations where you want to indicate the trend of a certain metric, the use of a baseline is very helpful to judge the insight. Let me quote an example from the video: imagine I want to indicate the forecast accuracy % (FCA%) of a company over time. Instead of creating a line chart with FCA% per month per year, it’s better to create a baseline of 100% compared in the same chart with the delta of the FCA% against that baseline:


Video and reference links

In the video below, I demonstrate the above best practices on landing pages, font types, color-coded KPI indicators and baseline graphs. They don’t cover everything you need to deliver staggering storyboards, but at least they give you a head start. In a next series of ‘Let Me Guide You’ blogs we will explore more tips and tricks.

Want to learn more on data visualization techniques? Check out:

Share on LinkedInShare on FacebookTweet about this on TwitterShare on Google+Email this to someone

This post is also available in: Dutch, French

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>