What are the key design principles and best practices for data visualization?

One way to think about is remembering the first time you got introduced to a new visual or idea. Did you feel overwhelmed? Make a concerted effort to keep it very simple is a tough job and takes a lot iterations to keep it simple. It’s better to have a curious audience rather than a confused audience. Start with simpler visuals which are more common to that set of audience. Researching your audience ahead of time is key. I often stick to bar charts, line charts and avoid anything else that would raise eyebrows

Some data visualization principles I go by: 1. On each frame/slide/view, less is more. Focus, trim, refocus, and trim again. 2. Main message is always front and center, never in a corner. 3. Apply colors sparingly and only to highlight the important content.

One simple concept that I learned but to this day I find it very insightful is to always consider how people usually look at a dashboard. That means to tell the "data story" from left to right and from top to bottom (a Z movement with the eyes). The less the person needs to think to understand the data visualization, the better the dashboard.

There is always a learning curve for the audience; it's not simply about the sophisticated or simple look. Sankey charts in one of my dashboards were initially quite unwelcoming to the users, but they are now its most popular feature. Hence, while clear data visualization is undoubtedly necessary, it shouldn't limit one's creativity or innovative thinking.

I recently got back from a seminar where I was one of many receiving recognition of the iCAS CSPA credential. One of the most influential exams was about data viz, and several of my new visuals were quickly adopted as company wide standards that all PMs used. If you don't like exams, just read the book from the exam on your own. It's called "Show Me The Numbers: Designing Tables and Graphs for Enlightenment" by Stephen Few. What I liked most about the book was that it didn't just help you understand the components of a good vs poor visualization. It also went into the biology and physiology of the eyes and how it works with short term memory so that you have a much clearer picture of why and what makes a good viz.

Research if the audience has in the past been usually more comfortable with utilizing static presentations like PowerPoint. If you are trying to change behaviors with users, you would have to make a more concerted effort to make that happen. As an example, for one of my dashboards, I tied the view to individuals employee id, so that it would in a way “force” individual users to navigate to the dashboard themselves and not delegating it to others

One thing that has helped massively in creating inclusive visualization is keeping in mind the color schemes we are working with. The conventional red, amber, green pallette might not be suitable for a diverse audience.

Tableau’s Viz Animation was something that in my opinion is a great example of innovation and creativity along with showcasing the simplicity and clarity of data if used and manuavered rightly. Experimenting with visualisations can many a times be overwhelming and could be defying by audience.

Data literacy is a very important factor for having a more distributed form of data culture. For one of my responsibilities with a recent organization, our goal was to have at-least one person on the team who can build a first draft of a dashboard with a tool like Tableau or PowerBI. Data governance and literacy are emerging as important concerns for organizations with huge amount of data but not enough value coming out of it