What are the pros and cons of interactive data visualizations?

In my experience, the drill-through and drill-down functionalities typically available in BI software like Power BI and the like are really powerful ways to quickly dissect the dimensions of your facts, e.g. '30% of revenue is coming from North East, within that region 75% is coming from x county and then you can see y and z locations are both the best generators of revenue'. I also think it's worth a bit shout out for AI visualisations available such as the Decomposition Tree, which is such an effective interactive visualisation to get quick insight.

flexibility and customisation should target or enhance visual thinking. This is important because humans have limitations when it comes to processing and focusing on large amounts of information simultaneous. However, we excel at identifying differences in visuals characteristics like length, positions and colour, hence flexibility and customisation should be key

Customization is a necessity for addressing the diverse needs of different audience groups. The ability to drill down, apply filters, and interact with granular data caters to those with specific interests while providing a quick overview that fulfills the requirements of others. Flexibility in customization is a valuable asset in meeting the varied data needs.

Just as there are multiple ways to unintentionally mislead with a static visual, there are also multiple ways to make things unintentionally not as insightful in an interactive visual. I opt for static visuals with annotations and careful color choices if I want to tell a story in a way that is easy to share and leads to a decision getting made. On the other hand, I opt for interactivity if the goal is more for exploration and an ongoing discussion. No matter which way I go, simplicity is key.

Key thing for me is be very very careful of understanding the filter context when designing interactive reports, the more complicated they are the more difficult it can be too keep track of all the different filters applying to visualisations, e.g. page filters, measure filters, visualisation filters. Ultimately, if you can deliver the message your stakeholders need to hear through a simple KPI card visual, leave it at that, don't try to add more than is necessary. There can be a tendency to worry about 'does this impressive enough?', but just go with what gets the message across and don't be precious over having loads of unnecessary, but beautiful nonetheless, visualisations.

A common trap I've fallen into when building interactive visualizations is the tendency to believe the audience will be as familiar with the underlying data as I am. As the author you are likely very familiar with the structure and logic of the data behind the visualization, but your users may be interacting with this data for the first time, putting them at risk of misunderstanding or making incorrect assumptions as they view the data, ultimately leading themselves to an incorrect conclusion. In my experience you can avoid this by giving your interactions a story like structure. Try and predict users followup questions from the initial view and then give them the tools they need to walk themselves through a self explanatory story.