Visualization is becoming an increasingly critical component of research studies that generate and analyze big data in biology, economics, physics, and many other disciplines. Visualization methods are sometimes categorized as “information visualization, which focuses on methods such as heat maps for showing high-dimensional research results, and scientific visualization, which focuses on the mathematics and physics of visualizing complex objects” [1]. There is a third emerging discipline called visual analytics, which integrates data analysis and human-computer interaction (HCI) with both information and scientific visualization methods. Data analysis methods such as those from machine learning and data mining have become much more powerful in recent years. HCI tools such as touch and gesture computing make it effortless to interact with visualization systems. These technologies are complemented by inexpensive 3D televisions and immersive display walls. Visual analytics is expected to take visualization to the next level by allowing researchers to explore their data in an interactive manner, while at the same time launching analyses from the visualization.

The ability to deliver visual analytics technology to researchers with big data will depend critically on the computational infrastructure that will support the visualization hardware. This paper presents a bulk synchronous visualization (BSV) method for interactive parallel computation. This is accomplished by breaking a big dataset into many small pieces that can be processed for visualization in parallel. Methods such as this will be critical if we are to deliver user-friendly visual analytics tools to big data researchers.

Visualization will only be effective when it is fast and seamless. Even small delays in rendering time can greatly impact the user experience and thus limit scientific discovery. This author provides one of many new methods that will be needed to fully exploit the patterns hidden in big data.