According to Stephen (2017), effective visualization meets a special set of criteria that address several features of effectiveness. A consistent set of criteria provides the means to compare the effectiveness of various visualizations and to discuss their relative merits with others using a common language. Shown together, how well a visualization achieves these criteria can serve as its data visualization effectiveness profile. The data visualization effectiveness profile that I’m suggesting consists of seven criteria, which fall into two general categories: criteria that address the degree to which visualization is informative (i.e., allows understanding) and criteria that address the degree to which it is emotive (i.e., presents a useful warm response). If we are thoughtful and reasonable, and we tell a great story, and we can present substantial experience to our viewers, that can mold our enterprise organizations in ways that just informing them will never do.

Hi,

Visualizations are usually and intentionally left open to interpretation so that various viewers can draw different interpretations explorable by getting into the data. Specific charts may not disclose certain important factors while they may be overstated in other charts which leads to much ambiguity. Data visualizations are often snapshot representations of a larger entity, and this can often lead to the creation of deceptive visualizations. Knaflic (2015) suggested generic guidelines applicable to most visualizations that  include:-

• Avoiding the use of 3D images as they skew numbers  and are a visual distraction
• Avoid the use of pie-charts as they are difficult to interpret especially when comparing a large number of variables
• Declutter the graphs to remove unnecessary visual elements
• Use color sparingly but consistently
• Provide underlying data to as replication and validation means
• All axes should be labelled in simple and understandable terms, and scales must be correctly illustrated

There is a global phenomenon on the shifting focus on end users with better, more useful data visualizations as business data visualization is an integral part of business intelligence systems. Creating effective visualizations is critical in the business data visualization as it supports decision making, analysis, and more importantly communication.

Regards,

Mohammed.

Reference:

Knaflic, C. (2015). Storytelling with Data: A data visualization guide for business professionals. New Jersey: John Wiley & Sons, Inc., pp.19-99.

The most successful visualizations are the ones that convey a message meaning in a few seconds. Coherence is vital, especially when creating visualizations from large data sets.  A coherent design should disappear into the background to enable users to process information quickly.

I would suggest that finding a balance between the number of data parameters, and how to visualize these is important : What type of feature controls what parameter, should the size represent a class of parameter (ex : for the parameter "cancer type" would be benign or malignant) ? should a color do that ? or maybe a shape ? Once a feature is assigned to a parameter, It's reserved uniquely for that parameter, else there would be an interpretation confusion. So I would guess that a higher dimension would make a visualization hard to design and maybe even harder to interpret. As Mohammed Kemal pointed out, A visualization needs to be intuitively and rapidly understood.

Hi Mohamed Benmahdjoub ,

Thank you for your reply, those are really great points. From my experience I want to share five principles for Great Data Visualization that I used it for many of my projects and want to share with you all:

1. Need relevant metrics

2. Make it visually attractive

3. Make it Interactive

4. Make it easy to use

5. Keep improving

What are your principles? Let me know your thoughts?

Mohammed Kemal Another principle would be to emphasize the key areas to help direct the audience' attention

First is the business value of your Viz

Then Context, is your data Viz coherent with each other

Interactivity and flexibility of filtering, sorting, exporting

Self-service capabilities, Allowing end user to extend or build their own dashboards

Layout adaptability to different display sizes and devices ( computer, tablets, smartphones...etc)

An excellent resource is the work of consultant and academic, Stephanie Evergreen. Her website has a huge amount of info, she regularly updates and sends newsletters too all for free.