Today we will cover such an invisible but important topic of standardisation in visualisation — we will find out why charts and dashboards should follow the same standard and how it makes work and decision-making easier.
Why Standardisation Matters?
Product Lead, Microsoft
Data visualisation standards are very important because they ensure consistency of views, which makes data clearer and easier to interpret.
It also helps designers to be more efficient because they don't have to spend time harmonising styles and formats. Standards also improve the perception of data as they make visualisations more consistent and understandable.
Visual Style Guide at Novartis
Data Analyst, EMEA
Typically, a style guide for teams involved in data analysis is perceived as something extra or unnecessary. Most likely, many would even agree that the quality of the data in the report and the high ROC-AUC of the predictive models is what end users really need, and let the designers handle your infographics.
The level of quality of visual information directly affects the accuracy of business decisions made by end-users.
In other words, poorly chosen charts and colours can just as easily lead to misleading insights as errors in the underlying data.
At the same time, the process of improving the look and feel of UX/UI best practices requires as much knowledge and practice as tuning models for machine learning. BI developers are usually at ease with heavyweight ETL, but struggle with choosing the right colours or graphics.
To combine the best of each world, we set out to create an actionable and easily scalable style guide.
It was important for us to create an application tool rather than a set of rules and guidelines.
The main objectives of the project looked like this:
Increase the speed of dashboard development
Create an easy to scale BI tool
Create a standardised view for all analytical reports
Enhance the user experience by increasing the quality and speed of business decision making by end users
Visual Vocab is a more practical way to learn how to create informative reports. It is an application with categorised visualisations developed for users at Novartis, most useful as a starting point in creating reports.
On the main page of the app, you can explore the description of categories, relate them to the tasks of the report, and then go to the necessary tab and select the type of diagram within the category to generate initial ideas about what may suit the purpose best.
The process of creating all elements of the style guide took 3 months, but it's hard to call it a finished project, more like a first edition. After the release, the tool continues to live and develop.
We regularly add and expand VizVocab'a functionality, refine formed BI practices that allow us to make development faster and user experience better.
How to Color your Dashboards
Incorrect use of colour in visualisations can completely ruin your work. The purpose of using colour can range from simple visual formatting, to drawing attention to and encoding additional variables.
The use of colour differs for different types of data, primarily for numeric measures and for specific categories.
Colour in diagrams and charts is a medium of information.
And if we have encoded information within a colour, and the colour is not obvious, it is difficult to read that information. Humans have a very strong correlation between colours and perceptions of the world — colour is supposed to support our patterns.
If the colour contradicts the patterns we have developed — for example, you used red on a chart "just for fun", then a person seeing the red bars will perceive the data as negative, i.e. bringing a negative effect. The user won’t even have time to look at the actual numbers that say otherwise. This is one of the ways of manipulation.
Make Your Own Dashboard
Continue the work started earlier and assemble a dashboard from the created diagrams.
→ Click here of you need help
If you are interested in learning more, follow the links to further explore the topics and expand your horizons.