Data Visualization in Data Science
Recommended by 0 users
“If you’re trying to extract useful information from an ever-increasing inflow of data, you’ll likely find visualization useful – whether it’s to show patterns or trends with graphics instead of mountains of text, or to try to explain complex issues to a nontechnical audience.” So writes InfoWorld’s Sharon Machlis.
Rebeckah Blewett, product manager for Dundas Data Visualization Inc., explains: “The practice of representing information visually is nothing new. Scientists, students, and analysts have been using data visualization for centuries to track everything from astrological phenomena to stock prices.” Data visualization, when done correctly, is a highly effective way to analyze large amounts of data to identify correlations, trends, outliers, patterns, and business conditions.
Many of us have experienced rudimentary forms of data visualization in our day-to-day experience of the Web. The popular TwitterEarth, for example, shows real-time tweets from all over the world on a 3D globe. It’s a great visualization tool to see where tweets are coming from in real time and discover new people to follow. It’s also fascinating just to sit and watch. Another simple example is the Flickr Related Tag Browser, which allows you to search for a series of tags and see related tags. Clicking on a different tag brings up new related tags. You can zoom into the tag selected in the center of the screen by hovering and see images tagged with that word. It also gives a total image count and lets you browse by page. And another is TED Sphere, which shows videos from the TED conference in a spherical format with 3D navigation. You can view the sphere from inside or outside and the layout of videos is based on semantic compatibility.
Data presentation can be beautiful, elegant and descriptive,” writes Vitaly Freidman of Smashing Magazine. “There is a variety of conventional ways to visualize data -tables, histograms, pie charts and bar graphs are being used every day, in every project and on every possible occasion. However, to convey a message … effectively, sometimes you need more than just a simple pie chart of your results. In fact, there are much better, profound, creative and absolutely fascinating ways to visualize data. Many of them might become ubiquitous in the next few years.”
In essence, the task of data visualization involves creating data layers and presenting these as easy-to-comprehend graphics for viewing by data analysts and non-tech decision-makers. Think of it as the graphical blending of data.
“Graph-based visual analysis is a highly effective method for capturing and understanding relationships between data that are not quantitative in nature,” writes industry pundit Jin H. Kim. “This method and technology has been used in diverse fields such as intelligence and law enforcement to customer sentiment and network topology analysis to uncover hidden insights in growing data that was not possible when relying only on traditional analytics.”
Kim continues: “The combination of rich data collection, advanced analytics operating across both structured and unstructured data, and efficiently storing and analyzing information in quantities unimagined just a few years back, have created a new era of data analysis in general and visual analysis in particular. We can now look at the networks representing relationships between data as not just static topologies, but rather as ‘dynamic networks’ with their own behavioral pattern in terms of change, sequence of change, and uncertainties of change, combined with the ability to integrate information from complex event processing engines and other ‘event driven’ information sources. These new developments promise to bring about a new dawn of information use, enabling smarter, timelier decision-making in various fields of human endeavor.”
As suggested previously (ala Twitter), data visualization plays a key role in real-time structured network analysis (SNA) – the modeling of relationships and overlaps between disparate groups of people.
“Social network analysis uses graph theoretic ideas and applies them with the premise that the structure of the graph can be used to understand and identify critical relationships and influential people. … ” writes Elizabeth Hefner of Tom Sawyer Software. ” Recent advancements in network analysis involving complex network topologies with multiple relationships between nodes, network behavior that is based on uncertain information, and time-based change of networks, have enhanced the value of incorporating advanced network analysis techniques as a key part of an analytics tool-set to aid in better understanding data relationships. More organizations are beginning to understand that with advanced visual analysis technology, they can build integrated insights across all of their available data, enabling them to better understand emerging opportunities and threats. … The combination of advanced visualization techniques, together with social network analysis techniques, will help bridge the emerging gap between the vast amounts of available information in Big Data and the available resources to better understand them.”
Numerous firms provide quite elegant, effective and powerful tools for data visualization.
These are just two options among many.
Edward Tufte – author of the classic Visual Display of Quantitative Information – has written: “The commonality between science and art is in trying to see profoundly – to develop strategies of seeing and showing.” No truer words.