What is data without visualization? Not much and certainly not very satisfying. From early cartography, tables, and data charts to dynamic dashboards, geospatial rendering, and network graphs, we seek tools to understand our world. Moore’s Law, web technology, mobile devices, economic storage and hi-resolution displays enable a revolution in information discovery and communication that seems only limited by our imagination. Fueled by distributed computing methods and machine learning techniques, data visualization of massive data sets serves to model and dramatically enhance our ability to discover and draw inferences about our world. Advanced graphical processors further extend the visual experience in rendering dimensional, virtual and augmented realities from data to produce immersive and augmented environments for further exploration.
"Data visualization as a mechanism for communication, is most effective when the “information story” is developed"
We use data visualization (DataViz) to reduce complexity, expand our perception and cognition, and to enhance our ability to communicate our ideas and insights to others. DataViz methods effectively represent the world through models and abstractions that permit exploration and experimentation. From data exploration we hope to develop insights into processes and interactions and to discover patterns and relationships that further our understanding of all things.
Experts have written extensively on the topic of effective data visualization. The guidance and recommendations provided by evangelists such as Edward Tufte, Stephen Few, Nathan Yau, and Alberto Cairo, have laid a solid foundation to confront our DataViz challenges. A quick search on YouTube for “Data Visualization” returns over 300,000 videos from individual practitioners, experts, products, and solutions to conference presentations and academic instructions. Numerous blogs by DataViz aficionados and experts alike are available online to provide the NeoData Scientists with guidance for tailoring the visualization experience.
DataViz seems to be as much about art as it is related to science as evidenced by the proliferation of infographics and applications available to develop creative visualizations. DataViz is pervasive in both print and web media and this has served to further raise the bar for succinct and effective messaging. Practically any DataViz effort is likely to pay significant regard to aesthetics and design in addition to the information communicated. DataViz is as much a right brain exercise as the left, presenting a challenge to create visualizations that effectively engage both sides to achieve utility.
The utility and effectiveness of DataViz for the most part depends first and foremost on the intended purpose. Exploring data to derive insights is quite different from creating a message addressed to an audience, such as a report or an infographic. Each DataViz opportunity has a specific objective and presents a different set of demands on your approach, selection of DataViz resources, tools, and constructs.
Data exploration requires all the data to be present in the discovery visualization, which is now computationally achievable at scale. High resolution displays enable our visual faculty to promptly discern patterns and relationships in large data sets that may not be readily evident in the large data tables or summary statistics. Data enrichment methods add to the discovery process to further enhance discovery such as applying geo-spatial reference or false color to reveal patterns. DataViz technologies provide the ability to add, subtract, combine, and derive data to further enhance our understanding and augment perceptions through direct programming languages to create the visualizations.
Data visualization as a mechanism for communication, is most effective when the “information story” is developed. While creating a report or an infographic, the data visualization ensures a clear and unbiased message through the viewer’s interpretation of the story. At times communicating information in a report may require the viewer to reason with and draw conclusions from the data and information presented. Sadly, it is feasible with today’s software packages to quickly develop tables and charts with numerous “bells and whistles” with little to no consideration of DataViz best practices. Report developers need to understand how the audience visually perceives information and draws inferences from the visuals. I find this to be a pervasive problem with business reporting in general. Financial data, Operational Business data, Human Resources and Information Technology all demonstrate poor use of charts including incorrect chart types and scales, missing data, color and more. In my opinion every organization should invest in training their report developers in the basics of data visualization.
In summary organizations use DataViz every day to transform data “facts” into the information necessary for informed decision making. Think about the medium and the message it conveys to the intended audience. Develop DataViz objectives to quickly, succinctly and effectively communicate the message. Focus on the quality of the DataViz and not the quantity. Reduce complexity, know the answer to the questions and communicate the story. As a personal check point ask the following about Data Visualization, is it effectively communicating, explaining and educating the viewer on what we intend to convey ?.